In Issue #143 of Better Bioeconomy, I want to explore a question that keeps surfacing as I read about AI-mediated purchasing: what happens when AI agents start filtering food choices before consumers see a shortlist?

Most commentary on AI and food retail focuses on the final act: the autonomous grocery order. That story is real, but it is not yet mainstream. The part that is already at scale, and underexamined, is the filtering step.

Before AI buys food for you, it is already deciding what gets considered. The shift underneath is structural: value is migrating from persuasion to proof, and legibility, the ability to produce machine-readable evidence of what you claim, is becoming the new shelf space.

The optimistic version of this story says quality finally gets rewarded because machines can see it. I think the direction is right. But the distributional question keeps pulling me back. The proof economy has a monetisation layer, a definitional politics, and a cost floor that together determine who participates and who gets excluded. That is what I want to sit with in this piece.

Let’s dig in!

AI filtering is already reshaping what gets considered

The share of US shoppers using AI assistants jumped from 12% to 35% in a single year (Adyen 2026 Retail Report, Censuswide survey, n=2,000). The definition is broad and captures everything from product research to checkout delegation, but even accounting for that breadth, the year-on-year move signals a behavioural shift that has already happened.

The pattern extends beyond the US. A Deloitte Asia Pacific report found nearly three-quarters of consumers in the region already use AI to discover and compare products, and 29% of APAC consumer businesses have adopted agentic AI, a figure projected to reach 76% within two years. At GrabX 2026, Grab launched an AI shopping agent that builds a grocery cart from a photo or voice note, with smart substitutions across merchants, no item-by-item browsing required.

Most of that activity is filtering, however, not directly buying. A 2025 Attest survey found 63.8% of consumers had used AI tools for food-related activities, with the majority concentrated in recipe recommendations and meal planning rather than checkout delegation. And when researchers at Yale, Columbia, and the University of Chicago tested what happens to products inside agent-mediated selection (featured in Kantar’s “What Is Your AI Agent Buying?” analysis), they found a 20-40% reduction in selection probability when a key product attribute was missing from the structured data.

Autonomous AI purchasing, meanwhile, remains structurally constrained. A Bluestone PIM survey from March 2026 found 73% of consumers uncomfortable with AI completing purchases autonomously, 68% assuming results are commercially influenced, and 65% refusing to store payment details. These are concerns about delegating spending authority and exposing household consumption patterns, the kind that do not go away just because the technology improves.

But filtering requires no delegation. The user asks a question, the agent answers, the user decides. The criteria are machine-readable or they are invisible. Packaging design, brand origin story, and shelf position, none of these register inside a filter that needs documented certifications, structured nutritional attributes, and composition data. That behaviour is already at scale, and it does not require autonomous purchasing to reshape which products get considered.

The Kantar figure is the one I keep returning to. A product losing 20-40% of its selection probability because its attributes are not machine-readable is facing a structural disadvantage. Which raises two questions. What determines whether a product is legible to the filter? And who controls the infrastructure that makes legibility possible?

When machines shortlist, value migrates from persuasion to proof

Here is a useful test for a food brand right now: if an agent filtered your product tomorrow, would it find machine-readable proof of what you claim, or would it find a story on your packaging?

If the answer is a story, the product is structurally disadvantaged in agent-mediated commerce. Better creative does not help a product score inside a filter that requires documented certifications, structured nutritional attributes, provenance chain records, and composition assays. The selection penalty that the Kantar study documents is what that disadvantage looks like in practice.

I find it useful to think about the infrastructure that processes these structured inputs in three layers.

At the top sit the agent platforms: OpenAI, Perplexity, Anthropic, Amazon Rufus, Walmart Sparky, and Instacart Cart Assistant. This is the interface layer, the surface where consumers interact, and shortlists get assembled. It is the layer most commentary focuses on because it is visible, named, and contested at high valuations.

At the bottom sits upstream measurement: soil sensing, nutrient assays, spectroscopy, and satellite imagery. This layer generates raw data about what is in the soil, what is in the crop, and what environmental and agronomic conditions shaped the harvest.

In the middle is what I am calling the claims-and-verification layer: the infrastructure that takes upstream measurements and translates them into machine-readable claims that agent platforms can query. This layer spans product content syndication, supply chain traceability, standards frameworks like GS1’s Digital Product Passport standard, and scoring systems that translate attribute data into structured outputs. These are different entities serving different functions, but they share a structural role: they manufacture the legibility signal that agents consume.

Antony Yousefian has argued the directional case well in The End of Brand Trust and The End of the Impulse Buy: verification costs collapsing, brand equity eroding, the farmer and soil layer reclaiming value. I agree with that direction. What I want to sit with here is where value concentrates inside the verification economy once it arrives, and who gets excluded from it.

My working view is that the translation step is where pricing power is most likely to sit, because it is the step neither the agent platforms nor the farms can fully own on their own. Platforms need verification but have incentives to commoditise it. Farms generate data, but most lack the aggregation infrastructure to translate it into agent-readable claims at scale, whether due to fragmentation, heterogeneous data formats, or the cost of standardisation across operations.

That is the clean version of the proof economy: quality that can prove itself gets rewarded, and the translation layer captures margin for making the proof legible. But three forces complicate the picture, and each one tilts the proof-equals-access equation in a different direction.

Three forces shift the proof-equals-access equation

1. Agent monetisation turns the shortlist into an advertising surface

The parallel I keep reaching for is SEO. In the early years of search, the dominant assumption was that good content, accurately indexed, would surface organically. Paid search gradually displaced organic results in most high-commercial-intent categories, and what emerged was a two-tier system where relevance determined eligibility and spend increasingly determined position within the eligible set.

Agent commerce is building a similar infrastructure. Stripe’s Shared Payment Tokens let agents pass scoped payment credentials to merchants. OpenAI introduced Instant Checkout in ChatGPT, then shifted toward directing shoppers to merchant storefronts, with Shopify building its own agentic storefront offering. Instacart’s Cart Assistant runs on the same commerce stack that underpins Instacart’s retail media business. Walmart Sparky now operates inside the Walmart app, ChatGPT, and Gemini, with checkout kept on Walmart’s own rails.

An agent monetised through retailer ad networks and platform revenue shares may not function as a neutral representative of the consumer’s dietary intent. Its shortlist would reflect some equilibrium between consumer preferences and platform economics, optimising for health within the constraints of whoever is paying for the infrastructure. That is a different dynamic from the rational health-optimiser that some of the more optimistic commentary assumes.

Machine-legibility gets a product into the eligible set. If agent commerce follows the search pattern, the position within that set will increasingly reflect spend.

Subscribe now

2. Whoever defines “healthy” controls the filter

When a user tells an agent “keep it healthy” or “low-inflammatory,” a translation has to happen. The natural-language instruction must map onto filter criteria applied to a product database. That mapping is unlikely to be neutral, and the entity that controls it shapes the outcome for the products downstream.

The FDA finalised an updated “healthy” nutrient content claim in December 2024, its first revision since 1994. It is a labelling standard. It is not a product-level scoring API that an agent can query in real time. When an agent operationalises “keep it healthy,” it queries whichever database has already done that translation work.

A handful of private systems have built that translation layer, and they do not all agree on what “healthy” means. Yuka scores products with nutritional quality weighted at 60%, additive content at 30%, and organic status at 10%. Open Food Facts uses Nutri-Score, scoring on a nutrient-only basis. HowGood is primarily a sustainability intelligence company, scoring on eight environmental and social metrics and integrating Nutri-Score separately for nutritional context. Personalised nutrition apps like Zoe score the same product differently for different users, calibrated to individual metabolic response.

Take a full-fat organic yoghurt with no additives. It scores well on Yuka (nutrition plus clean ingredients), variably on Zoe (depends on the user’s metabolic profile), and differently again on HowGood (depends on the sustainability footprint of its supply chain). An agent shortlist filtered for “healthy” produces different results depending on which system it queries, and the consumer has no visibility into which taxonomy shaped the shortlist.

The analogy that keeps pulling me is pharmacy formulary placement. In pharmaceutical distribution, a drug’s presence on a formulary often determines patient access more directly than its clinical profile. The formulary decision happens upstream of the prescription. In agent-mediated food commerce, the taxonomy endorsement happens upstream of the shortlist. A product’s presence in the scoring system of a given platform's queries may matter more to its agent-era visibility than its placement on any retailer’s digital shelf.

3. Platform fragmentation multiplies the legibility target

The optimistic version of this transition assumes interoperability: one proof layer, universal reach. That assumption is not holding up.

Last month, U.S. District Judge Maxine Chesney granted Amazon a preliminary injunction blocking Perplexity’s Comet browser agent from accessing password-protected areas of Amazon, including logged-in Prime experiences. A brand’s presence inside an authenticated Amazon shopping flow, the one with rich personalisation data, is not addressable by agents operating outside Amazon’s perimeter. The emerging protocol standards reinforce the pattern. OpenAI’s Agentic Commerce Protocol and Google’s Agent Payments Protocol (AP2) each require separate integration work, and neither is emerging as a universal standard.

In Southeast Asia, Google and Sea Limited announced a partnership to build an agentic shopping prototype for Shopee, which holds 52% of the region’s e-commerce market, and to pilot agentic payments through AP2 via Sea’s fintech arm Monee. That is yet another platform-specific integration a food brand would need to reach agent-mediated shoppers in a different region.

For a mid-sized food brand trying to be legible to agents, this means the target is not one data standard but several, each with its own integration cost and taxonomy. Machine-legibility is not a single investment. It is one investment per platform that maintains a walled experience layer.

Fragmentation also resolves a structural question I raised earlier: whether the claims-and-verification layer retains independent leverage or gets absorbed by platforms. If every platform maintained the same taxonomy, a single acquisition could internalise the middle layer’s margin. In a fragmented environment, no single acquisition gives a platform full coverage. The middle layer retains leverage because it is the translation infrastructure between incompatible systems. The more fragmented the protocol landscape stays, the more durable the middle layer’s independent pricing power.

Taken together, these three forces mean the legibility economy is not a clean meritocracy where proof equals access. Access depends on proof, position within the eligible set reflects spend, and the definition of proof depends on which taxonomy the agent queries.

The proof economy’s cost floor determines who participates

Even setting aside monetisation, definitional politics, and fragmentation, the basic economics of producing machine-verifiable evidence sort food producers by scale.

Certifications, provenance chain documentation, composition assays, soil health records, nutrient density measurements: these are the attributes that move a product from invisible to eligible in agent-mediated shortlists. Supplying verifiable evidence is the new table stakes for commercial relevance.

The unit economics do not distribute evenly. A soil assay at agronomic resolution has a cost. Spectroscopy-based nutrient analysis at the lot or batch level has a cost. Third-party certification audits also have a cost. Building the data systems that retailer traceability mandates and GFSI-level food safety programs increasingly require its own investment.

The EU’s Ecodesign for Sustainable Products Regulation (ESPR), which introduces Digital Product Passports for priority sectors like textiles and batteries first, signals where food and agricultural products are likely headed next. At scale, the per-unit cost of all this is manageable. At five acres, it may exceed the margin the producer is working with.

The consequence is a sorting mechanism embedded in the proof economy’s structure. Mid-sized producers with the operational capacity to implement sensing infrastructure and the volume to amortise it gain access to agent-mediated shortlists. Artisanal and small-scale producers, the kind of long-tail supply that specialty food brands and farm-to-table operators have built sourcing differentiation around, face a proof layer that is technically accessible but economically out of reach at current cost structures.

Where this thesis might be ahead of itself

If sensing costs follow the trajectory of other AI-adjacent hardware, falling significantly per cycle as compute and miniaturisation improve, the cost floor argument may be temporary. The distributional effects I describe could be a transitional feature of early infrastructure rather than a permanent structural condition.

If platforms consolidate around a single protocol, with GS1’s Digital Product Passport being the most plausible candidate, fragmentation collapses and the middle layer’s pricing power weakens. A unified standard would simplify legibility into a single investment rather than one per platform, which is better for producers but changes the value map I have outlined.

And if agent monetisation evolves toward subscription models rather than advertising, the shortlist may stay closer to consumer intent than the SEO parallel suggests. The search pattern is one possible outcome, and it may not be the dominant one.

The direction of the shift, from persuasion to proof, is clear to me. The speed, the severity of the distributional effects, and whether the cost floor proves durable or temporary are the open questions.

What this means for you

Every food brand needs a legibility audit before agents run one for them

The diagnostic question I posed earlier in this piece, whether an agent would find machine-readable proof or a story on your packaging, is worth turning into an exercise.

Work through it: what structured nutritional attributes does your product have in queryable databases? Which scoring taxonomies (Yuka, HowGood, Nutri-Score) can access those attributes? Which agent platforms can see your product data inside authenticated environments? This is not a five-year planning horizon. The filtering is already happening.

The most defensible infrastructure plays sit in multi-protocol translation

For investors evaluating data infrastructure in this space, the fragmentation argument implies a specific thesis. The most defensible companies are those whose value increases with fragmentation: multi-protocol translation capability, cross-platform data syndication, taxonomy interoperability.

These companies survive consolidation (they become the acquired integration layer) and thrive during fragmentation (they are the only path to multi-platform legibility). This is a different filter from “invest in the agent platforms” or “invest in the sensing layer,” both of which face commoditisation pressure from different directions.

Whether legibility becomes a public utility determines who participates

The cost floor and the definitional politics I have described will not self-correct through market dynamics alone. Three interventions would change the distributional math: cooperative data infrastructure that lets small producers pool sensing and certification costs across shared operations, public scoring standards that prevent private taxonomies from becoming the sole gatekeepers of “healthy” or “sustainable,” and protocol-level interoperability mandates that reduce the per-platform cost of legibility.

These are policy choices, and they need to be made before the architecture hardens. The sector needs to decide whether legibility is a public utility or a private good. That decision will shape whether the proof economy creates a more open food system or reproduces the concentration dynamics of the shelf-space era in a new form.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it!

Share

Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Disclaimer: The views and opinions expressed in this newsletter are my own and do not reflect those of my employer, affiliates, or any organisations I am associated with.

Welcome to Issue #142 of Better Bioeconomy, thanks for being here!

Agritech funding in Southeast and South Asia (SEASA) peaked at more than $750 million in 2022, then dropped roughly 70% by 2025 in both deal count and funding volume. Most people read that as a market failing.

But reading the recent report, “The Opportunity for AgriTech Investment in Southeast and South Asia” (published by Beanstalk AgTech and Briter Bridges, commissioned by Omnivore, FMO, IFC, and Rabo Foundation), I kept arriving at a different conclusion. The opportunity metrics haven't moved. What pulled back was a specific capital model that was not built for how value gets created in these markets.

That difference between a market correcting itself and a capital model revealing its own structural limits is what I want to explore. Because what SEASA is teaching us about capital architecture in emerging agricultural markets is relevant beyond the region.

Let’s dig in!

The funding dropped 70%, but the opportunity metrics didn’t move

The report covers the SEASA region: Southeast Asia (Vietnam, Indonesia, Thailand, Philippines, Myanmar, Cambodia) and South Asia (India, Pakistan, Bangladesh, Sri Lanka), drawing on deal-level data, crop-specific value chain analysis, and a mapping of the startup and investor ecosystem.

Agriculture contributes ~15% of GDP and employs 30-42% of the workforce across SEASA. Between 2020-2022, generalist VCs poured capital into the region. When funding collapsed by 2025, the pullback was sharp and broad. But the structural drivers of the opportunity stayed put.

Yield gaps, post-harvest losses, and SME financing shortfalls remain largely unchanged. The report estimates that digitalisation and agritech adoption could unlock more than $90 billion in annual GDP gains in Southeast Asia alone by 2033.

The report makes another observation I can attest to: agritech rarely ranks as a priority sector for venture capital allocation. It sits at the bottom of what the authors call the “investment waterfall,” with capital flowing first to climate tech, fintech, and broader impact sectors. In my experience, even within climate tech across the region, most allocator attention goes to energy and transportation, with agrifood treated as an afterthought despite its outsized impact potential.

So the capital that flooded in during 2020-2022 was not just the wrong structure. For many funds, agritech was never the core conviction. The 70% funding drop was a failure of the capital model on its own terms, not a market failure.

SEASA is a collection of distinct, localised markets, each with its own rules

The report clearly maps the fragmentation of the region. SEASA is not a market. It is a collection of distinct, localised markets, each with its own policy regimes, value chain dynamics, and consumer patterns.

The report draws a foundational distinction between two types of agricultural markets that coexist across the region. Smallholder-led staple value chains, rice, maize, and wheat in Bangladesh, Cambodia, and Nepal, are high volume, low margin, and oriented toward domestic food security. Highly commercialised, tech-adopting chains, rubber, coffee, and palm oil in Thailand, Vietnam, and Indonesia, are export-oriented, compliance-driven, and increasingly venture-legible. Same region, different investment theses.

The fragmentation runs deeper than that divide. Even within a single crop, where value concentrates in the supply chain varies by country. In Vietnamese shrimp farming, profits split roughly evenly between upstream production (51%) and downstream processing (43.5%).

In Philippine coconut production, 74% of the profit pool sits upstream in production and primary processing. The point is not to compare shrimp with coconut. It is that the investment thesis has to be crop-specific and country-specific. Backing a downstream aggregation play without understanding where the profit sits in that specific value chain could lead to misallocation of capital.

Premature regional expansion was one of the most common failure modes. Of the entire SEASA ecosystem, only 12 startups attempted cross-border scaling, with approximately two-thirds failing or pulling back. This matches what I’ve seen working in the region. The pitch deck says “Southeast Asia,” but moving from Indonesia to Thailand could mean rebuilding distribution, regulatory relationships, and tweaking the product itself.

In SEASA, the funding that arrived in the first wave was structured for regional scale. The assumption was that a model proven in Vietnam could be replciated into Indonesia, then the Philippines. The markets disagreed.

Strong local traction, failed regional expansion

That fragmentation created the conditions for a specific and repeatable failure pattern. The vast majority of SEASA startups surveyed operate at negative margins. By 2022-2023, 69% of the $569 million in funding went to startups with demonstrated unit profitability, suggesting investors had already learned to screen for it.

The report identifies what it calls the "Series C Trap." Early-stage companies achieve strong local growth by solving a specific friction point in a single market. Earlier rounds up to Series A can generate real traction. The trap springs later: under pressure to demonstrate the scale that justifies a Series C valuation, companies attempt cross-border expansion into markets with fundamentally different dynamics. Distribution costs can consume 30-40% of revenue in fragmented SEASA markets, and the result of premature expansion is diluted product-market fit and eventual failure.

What the first wave exposed: unit economics matter more than the total addressable market. The region has 71 million farmers, but only ~5% use digital platforms with an average spend of below $50 per year. The addressable market was a fraction of what pitch decks suggested. Direct-to-farmer models made it worse: customer acquisition costs ran 3-5x higher than B2B or intermediary approaches, and even after acquisition, serving smallholders at that price point was economically challenging. Onboarding, support, and logistics costs eroded whatever margin remained.

Over 60% of failures between 2022-2025 cited poorly sequenced scaling as the cause. The companies that survived share a pattern: they dominated a single market before expanding, built local execution teams, and treated distribution as a core competence rather than a fundraising milestone.

The typical Series C timeline assumes that the infrastructure for scaling already exists: reliable distribution, functioning logistics, and available financing for customers. In SEASA, that infrastructure often does not exist. The VC scaling cadence collided with markets that require ecosystem-building before, or at least alongside, company-building. There's a related dynamic I see often: VC pressure makes startups feel they need to ramp faster than their demand requires, leading to CAPEX investments in capacity that sits underutilised. The capital model doesn't just fail at the wrong stage. It can distort operating decisions well before that stage arrives.

Three demand drivers that exist regardless of VC sentiment

Despite everything the first wave got wrong, the structural drivers of the opportunity were never the problem. Reading across the report's value chain analysis, three pillars stand out, each grounded in constraints that technology can address, and each growing independently of venture capital sentiment.

What connects all three is that they are pull markets: the demand exists independently of whether a startup shows up to serve it.

Productivity gaps and the $40 billion in annual losses

Production yields across SEASA remain 2-6x lower than the highest-yielding global producers, and post-harvest losses add $40 billion in annual drag from inadequate cold storage and fragmented logistics. But not all of that gap is the same kind of problem.

Countries with high Total Factor Productivity (TFP) growth, like Vietnam, are modernising: output is rising while the labour index is declining, signalling a sector where labour is becoming more productive. The technology opportunity here is precision agriculture, input efficiency, and automation.

Countries with low TFP growth, like Pakistan, are growing through extensification: output gains are matched by increases in labour and land inputs. The opportunity there is more foundational: basic mechanisation, improved inputs, and knowledge dissemination platforms.

This matters for investors because it means the same sector label (”productivity tech”) describes two completely different investment theses depending on the country.

Premiumization and compliance as a paid market

Beyond yield gains, value is increasingly created through premiumization, particularly for high-value export crops like coffee and palm oil. The rise of mandatory compliance pressures, most notably the EU Deforestation Regulation (EUDR), has turned what was once an optional ESG upgrade into a non-discretionary requirement.

Technology solutions for MRV (Measurement, Reporting, and Verification, the tools that track, measure, and prove environmental claims) and traceability are creating what is effectively a paid market for compliance. B2B SaaS platforms that bundle compliance software with verification services can capture value through mandatory traceability adoption. The scale of the premium can be seen in specialty categories already: coconut sugar commands 1,600-5,000% premiums, and premium seafood 650-20,000% over commodity equivalents.

When a regulation requires compliance, the willingness-to-pay question largely disappears. But the open question is who captures that value. If the tech provider is enabling a premium that accrues to the exporter, the SaaS fee has to be priced against margin the exporter would not otherwise have. That is a strong position, but only if switching costs are high enough to prevent the exporter from building the capability in-house once the compliance workflow is proven.

Climate adaptation is creating “market pull” for non-discretionary tech spending

Drought, salinity, cyclones, and extreme heat impose unavoidable adaptation costs across SEASA agriculture. The report frames this as a “pull” market: climate technologies are shifting from optional to essential expenditure.

What makes this useful for investors is that the climate patterns themselves dictate distinct technology priorities by geography. Drought-resistant genetics and water optimisation (like Alternate Wetting and Drying for rice) map to mainland Southeast Asia and inland South Asia, where water stress is most acute. Salt-tolerant crop varieties target coastal regions facing dual threats of flooding and salinity from sea-level rise. Storm-resilient infrastructure and parametric insurance concentrate along exposed coastlines in the Philippines, Vietnam, and eastern India, where cyclone vulnerability is highest.

Each of these is a different product category, a different customer, and a different go-to-market. The climate thesis in SEASA is not one investment. It is at least four, each geographically constrained.

That said, policy design matters as much as climate exposure. Vietnam’s One Million Hectare rice program has successfully catalysed smart cultivation and traceability pilots at national scale. Pakistan’s biofertilizer startups face headwinds from entrenched chemical subsidy regimes. Same regional challenge, opposite policy environments, very different adoption curves.

The harder question remains adoption velocity: parametric insurance has been “gaining traction” for years, but actual penetration among smallholders remains thin. How fast structural demand converts into revenue depends on distribution infrastructure that, in much of SEASA, does not yet exist. This is the same constraint that doomed direct-to-farmer platforms in the first wave. The opportunity is real. The business model that can access it without building infrastructure from scratch is what separates a viable business from an expensive lesson.

Subscribe now

Three verticals where the economics work

Not all SEASA agritech verticals are waiting for the same kind of capital to come back. What connects the verticals showing signs of life is strong underlying economics. What separates them is the capital structure each one requires.

Aquaculture and marine innovation

Aquaculture contributes $26.9 billion annually to Indonesia’s economy alone and employs millions across Vietnam, India, and Bangladesh. This is the vertical where the gap between potential and current utilisation is widest. Indonesia’s seaweed industry alone could reach $11.8 billion by 2030, yet only 0.8% of suitable hectares are currently utilised.

Feed accounts for up to 57% of total costs for shrimp and 84% for pangasius (a freshwater catfish native to Southeast Asia), so any efficiency gain in feed conversion translates directly to margin. Platforms like Tepbac (Vietnam) and AquaEasy (Singapore) claim feed savings of 20-30% through AI-driven farm management.

The sector is also adapting its business model in response to first-wave lessons. 61% of aquaculture startups now offer bundled solutions combining hardware, software, inputs, and services, with marketplace-plus-SaaS platforms attracting the largest rounds.

Despite high-profile failures in the sector, the demand for aquaculture protein continues to grow, and the production inefficiencies remain massive. But the capital this vertical needs likely looks more like project finance than venture equity: long-horizon, asset-backed, tied to infrastructure buildout.

Inclusive agri-fintech

The region has a $12-15 billion SME financing gap, and the models that are gaining traction have stopped trying to build distribution from scratch. New approaches are shifting from prohibitive “direct-to-farmer” lending toward trade-data underwriting, where transaction histories with established distributors become the basis for credit decisions.

By partnering with existing distribution networks rather than building their own, these platforms can achieve unit profitability through transaction fees and data monetisation. A related growth area is parametric insurance for aquaculture and livestock, which the report identifies as a billion-dollar opportunity, building on the climate adaptation demand outlined above.

The test for whether an emerging market agritech model will work is whether it can reach customers through infrastructure that already exists, or whether it needs to build the infrastructure first. The models gaining traction in SEASA are built on what is already there.

Of the three verticals here, agri-fintech is probably the one most naturally suited to standard venture capital: it runs on existing distribution rails, generates transaction-level data from day one, and can reach unit profitability without building physical infrastructure.

Sustainable feed and biotech

Sustainable inputs like algae and insect protein are attracting investment driven by clear demand from feed mills across the region. In the Philippines, aquaculture biotech ventures are integrating upstream solutions (the biology) with downstream brand partnerships (the market access).

With feed as the dominant cost driver across aquaculture species, alternative protein inputs are being pulled by the sector’s own cost structure. This is demand-led innovation, not supply-pushed technology looking for a market. And the capital structure follows the demand: corporate VCs from feed mills and processors (Thai Union Ventures, CP Group, Wilmar) are the natural funders here, because they are also the customers. This is perhaps the vertical that has the strongest set of acquirers.

Corporate acquisitions at $200-400M are the market revealing what works

If the first wave’s capital model was wrong, what does the right one look like? The exit data is a good place to start, because exit architecture tells you what the market rewards.

Since 2020, corporate acquisitions have accounted for roughly three-quarters of all SEASA agritech exits. Only eight IPOs have been completed across the broader ecosystem. The realistic exit target is $200-400 million, driven primarily by strategic buyers like Wilmar, Thai Union, and CP Group looking to de-risk supply chains and accelerate digital transformation.

This tracks with a pattern I covered in Issue #137 on the Verdant M&A review: thesis-driven corporate buyers are replacing conglomerate logic as the primary acquirers in food and agriculture globally. In SEASA, the same dynamic plays out with regional corporates filling specific platform gaps.

And exit proceeds are rarely recycled into new funds, creating what the report calls a “single-use” dynamic: without successful exits cycling capital back, new money must constantly flow in, compounding liquidity constraints across the ecosystem.

Niche exit pathways are also emerging. In India, SME listing platforms like the Bombay Stock Exchange SME have provided liquidity for smaller ventures with post-issue paid-up capital as low as $3-$4 million.

From my experience in the region, interest in secondary transactions is also picking up as early investors look for liquidity without waiting for a full exit. This is a global trend: as Cyril Ebersweiler, General Partner at SOSV, a global deep tech venture firm, shared in our conversation, secondaries grew from 10% to 28% of exit volume globally between 2020 and 2024, capturing 71% of exit dollars in 2024. If that trend reaches SEASA, it could ease the “single-use” capital constraint that currently starves the ecosystem of recycled capital.

Blended finance is the right architecture, with caveats

On the capital side, the architecture replacing VC primacy is blended finance: equity combined with credit and concessional capital. DFIs like the IFC, FMO, and Rabo Foundation are emerging as critical anchors, acting as cornerstone LPs in most newly launched agrifood funds with a combined $650 million committed to date.

Local financial institutions now participate in over a quarter of corporate-backed agritech deals, blending equity with working capital or ecosystem lending. Corporate VCs are the second-leading investor type after traditional VCs, providing both capital and the market validation that a business solves a real problem for a buyer.

Companies that have revenue profiles supporting debt instruments are increasingly raising capital through non-equity channels. The capital stack is diversifying because it has to. As Oscar Ramos, Managing General Partner at Orbit Ventures, an emerging-markets venture platform, mentioned in our conversation, designing a capital stack that matches the realities of the business is becoming just as important as designing the product itself.

A note of transparency: this report was commissioned by Omnivore, FMO, IFC, and Rabo Foundation, all of whom operate within the blended finance and impact investing model the report advocates. That doesn’t invalidate the analysis, the data is rigorous, and the findings track with what I see in practice. But you should hold the report’s enthusiasm for blended finance with an awareness that its commissioners have a structural interest in that architecture succeeding.

And on that point, while I am all for blended finance, I want to caution that blended finance is not as clean a solution as it sounds on paper. The trade-offs are real, and founders need to understand them.

The first is speed. DFI capital comes with decision timelines that can stretch months longer than commercial rounds. A startup competing for a distribution partnership or reacting to a regulatory window may not have that time.

The second is flexibility. Impact mandates can directly conflict with the pivots that startups need to make to survive. Consider an alternative protein startup backed by concessional capital tied to plant-based or fermentation-derived outcomes. If that company discovers its fastest path to unit economics runs through a partnership with an animal-based protein processor (co-manufacturing, shared cold chain, blended product lines), that pivot may be incompatible with the terms of its funding.

Blended finance is the right structural direction for SEASA. But the cost is flexibility and speed, and in markets where conditions shift fast, those can add up.

What I’m taking away from this

Build for acquisition readiness from the start

In a market where three-quarters of exits are corporate acquisitions in the $200 to $400 million range, the exit strategy is a founding thesis decision, not a Series B conversation.

The diagnostic questions for founders: Which three to five corporates are the most likely acquirers for your specific capability? Are you building something that slots into their workflows, or something that requires them to change how they operate?

And critically: don’t “go regional” until your local ecosystem can sustain the growth. The Series C Trap is a sequencing problem, and the sequencing starts at founding.

Fund architecture must match exit architecture

If the exit ceiling is $200 to $400 million, the fund math dictates everything downstream. A fund sized in that range can underwrite to these exits and generate meaningful returns at reasonable ownership levels.

A $200 million-plus fund requires a fundamentally different portfolio construction to make the math work, perhaps through multi-stage reserves, or a hybrid equity-credit model. The question for LPs: are fund sizes being calibrated to the market's exit architecture, or to LP appetite for larger commitments?

For investors building emerging market theses beyond SEASA

The structural conditions that make standard VC architecture a poor fit for SEASA (fragmented local markets, smallholder dominance, compliance-driven demand, infrastructure gaps that precede software adoption) exist in Sub-Saharan Africa, parts of Latin America, and Central Asia.

The SEASA correction is a dataset on what happens when capital architecture is imported rather than designed locally. The fund structures, check sizes, and exit expectations that emerge over the next two to three years will become templates, both for what to replicate and what to avoid.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it!

Share

Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Disclaimer: The views and opinions expressed in this newsletter are my own and do not reflect those of my employer, affiliates, or any organisations I am associated with.

Thanks for being here. For Issue #141 of Better Bioeconomy, I sat down with Steven Finn, Co‑Founder and General Partner at Siddhi Capital.

Siddhi is an investment firm operating across two distinct verticals: consumer packaged goods brands and food technology. What makes the fund structurally interesting is that more than half the team aren’t investment professionals. They’re operators, inherited from a food systems consulting firm co-founded by Steven’s partner Melissa.

In 2022, that firm was folded as a standalone business, and its people were brought into Siddhi’s management company to work directly with the portfolio. The result is a fund that can, as Steven puts it, build your food facility and write you a check.

Steven leads the food tech side of the house. He came to it as an engineer first and an investor second, which shapes how he thinks about B2B businesses: always global in scope, always grounded in commercialisation mechanics. He’s refreshingly outspoken about the mistakes that got the sector into its current mess, and equally clear about where he sees the path out.

In our chat, we spoke about why the food tech investing climate is bad (and why that’s not the whole story), the structural trap of trying to do CPG and deep tech under one roof, how to think about cap tables in a sector still recovering from its own “peak stupid”, and where GLP-1 and AI are creating real opportunity versus just noise.

Let’s jump in!

The investing climate is bad, but drivers for innovation didn’t go away

The food tech investing climate is bad. Steven won’t sugarcoat it. But he draws a line between the sector’s PR problem and the underlying quality of companies being built right now.

The issue is structural to private markets. When a public market sector crashes, it crashes together, a shared bad moment that the market can process, price in, and move on from. Private markets don’t work that way. The companies that raised too much money at wrong valuations during what Steven calls “peak stupid,” roughly 2019-2022, didn’t fail at the same time. They’ve been dropping one by one over the years. That creates a bad news cycle that persists long after the underlying cause has passed.

The consequence: generalist money, which flooded the sector during peak stupid and then got burned, has stayed away. Access to capital has dried up. Valuations have compressed. And the slow drip of high-profile failures makes it look like the whole space is broken, even when companies founded later are doing compelling work on more disciplined capital structures.

“Dumb generalist money went to the wrong places in the first place,” he said. “That’s not because the companies aren’t doing great, amazing things. It’s because they raised the wrong amount of money at the wrong time.”

What he thinks people are missing is that the drivers for food tech didn’t go anywhere. Supply shocks are still coming. Wars, fertiliser disruptions and geopolitical fractures keep reminding the world how fragile food supply chains are, and each new crisis pushes eggs, cocoa, oils, and other categories back into the spotlight.

The difference this time, he argues, is that the biotech and ingredient infrastructure is “many steps closer” to being able to respond than it was during Covid. The tools are better, the playbooks are clearer, and there are more teams building around real price and taste rather than purely around narratives.

Siddhi’s answer is to keep writing checks, but not into companies that require a US$100M generalist-led round in 12 months to survive. Instead, they back businesses that can raise smaller, staged amounts from industry insiders while the market heals and can extend runway without dropping two-thirds of a round into CAPEX on day one.

The bet is that by the time these companies are ready for bigger capital, a couple of visible wins will have done the work of bringing generalist money back in. Not on 2021 terms, but on terms that reflect what the sector has learned.

Don’t try to do ‘tech and brand’ under one roof

Siddhi invests in both pure CPG brands and deep tech ingredient platforms, but Steven is adamant about one line they won’t cross: they do “zero companies that are trying to do both.” The two tracks run separately, collaborate closely, and don’t merge.

That conviction is shaped by scars from the consumer side. After years of getting “persistently and occasionally beaten” in CPG and watching brilliant scientific founders treat the consumer brand as the “easy part,” he has come to see tech and brand as two separate zero-to-one problems.

Putting both zero-to-one journeys under one roof does not double complexity. It turns it into an order-of-magnitude problem. The teams are different, the skill sets are different, and, crucially, the investor universes and valuation logics are different. For a tech platform, the early valuation may be anchored on IP, platform potential and B2B unit economics. But for a consumer brand, valuations are often anchored to revenue multiples and gross margins.

When a deep tech company bolts on a small brand because “the industry is moving too slow to adopt us,” it typically ends up being valued like a consumer business doing a few hundred thousand dollars in sales rather than like an ingredient platform with scalable upside.

“What was your last valuation?” he asks ingredient founders considering a consumer pivot. “Because the second you’re a consumer product, you’re valued as a consumer product, and you’re probably worth two to four times sales.” Go out and sell $100,000 of product in your first year, and you’ve just marked yourself down to $200,000 in enterprise value, regardless of what the underlying science is worth.

The bifurcated model Siddhi runs creates an edge on both sides. The CPG portfolio, growth-stage brands doing roughly $30-150 million in revenue, are big enough to be meaningfully distributed but small enough to move quickly and run ingredient pilots at small quantities. That gives the food tech side a real-world commercialisation channel and live feedback on what problems the CPG market needs solving. One side feeds the other. But only because they stay separate.

A great consumer brand is more defensible than a patent

Most investors in this space instinctively favour the patent. Steven doesn’t. His view is that both a brand and a patent are essentially zero-to-one exercises, rare, hard to build, and not as durable as they appear once achieved.

What undermines a patent in food specifically is that customers don’t care about the method. They care about the outcome. And in food systems, there is almost always more than one way to achieve the same outcome.

Take a cultivated meat bioreactor company, which builds a compelling case: there aren’t enough steel bioreactors in the world to feed people at scale, it’s a real constraint. That argument resonates until the day you receive a pitch on molecular farming, where plants become the bioreactors, and the capital intensity problem largely disappears.

Same consumer goal. Entirely different path. “We’re solving the same problem so many different ways,” Steven said, “that I think patents are great, but businesses are better.” The clearest implication: IP is a starting position, not a finish line.

What this means practically is that Siddhi evaluates defensibility less around IP and more around whether a company has a real business: customers, recurring revenue, and relationships that can’t easily be replicated.

A brand that has earned shelf space, consumer loyalty, and retailer trust is harder to dislodge than a patent that a well-funded competitor can design around by approaching the same problem from a different scientific angle.

Subscribe now

The “financing risk first” mindset to underwrite startups

The phrase “financing risk first” is Steven’s shorthand for how Siddhi underwrites deals now versus four or five years ago. The key question he asks early is: “How many dollars do you need before you don’t need any more dollars?” From there, he works backwards into a multi-year fundraising plan tied to specific value inflection points and milestones that change the way a future investor can underwrite the business.

In the “peak stupid” years, founders could get away with a hand‑wavey “we’ll raise a big Series B from generalists in 12 months” as part of the story. Today, that is an immediate red flag. Siddhi wants to see rounds sized and timed around realistic timelines to tech de-risking and commercial proof.

That includes a path to strong EBITDA, not just the maximum cheque size the market might bear. And it includes practical questions: can this company extend runway if needed, or have they structured the business so that two-thirds of a round goes into irreversible CAPEX on day two, leaving no ability to slow burn or pivot if the next raise takes longer than expected.

This financing‑risk lens also shapes which stage Siddhi is solving for in which vehicle. Their earlier-stage capital is often aimed simply at getting a company to a credible Series A: proof that the tech works, first customers, a sense that there is a real market pull.

The main growth fund, which writes larger checks and leads Series A and B rounds, underwrites an exit case that looks more like private equity than IPO. That means baking in the assumption that the business will need to become highly profitable and will likely be sold on EBITDA multiples, with strategic acquisitions treated as upside rather than the default exit path.

In practice, for a growth-stage entry, they are aiming for a world where, if things go right, their base case looks like a 5-6x return on invested capital, which they then mentally haircut to a 3x to account for the fact that “anything that can go wrong will go wrong.” That math only works if the company is still alive at the exit. Which is why the conversation about unit economics, total capital required to reach cashflow positivity, starts on day one.

A cap table can kill a good company

Steven is almost evangelical about cap table engineering. He has written multi‑part LinkedIn posts breaking down how preference stacks and misaligned valuations quietly kill otherwise good food companies, and those posts are how I first came across his work.

For context, a cap table (short for capitalisation table) records who owns equity in a company and on what terms. In any fundraise or exit, it determines who gets paid first and how much.

His diagnosis of the last cycle is that many companies raised the wrong amount of money at the wrong time, at the wrong price, and then felt compelled to take only “huge swings” to try to justify it. As those swings missed, as most huge swings do, the cap tables that had once signalled ambition turned into anchors.

If a company raised US$100M at a US$300M pre‑money valuation, and now needs to raise at US$5M pre because the world moved on, a naïve approach might be to just stack the new round on top and hope for the best. In reality, that leaves US$100M of liquidation preference ahead of the new investor, which means most realistic exits go almost entirely to the old money and leave the new round with very limited upside.

That kind of structure kills deals: new investors cannot underwrite to a reasonable outcome, teams know they will never see meaningful upside, and everyone is trapped waiting for a fantasy exit that will never come.

Steven’s rule is that the cap table must reflect the moment the company is in and the path it is now on, not the frothiest part of the last cycle when a prior round was raised. If the business model, market, or financing reality has changed, the cap table has to change too. Otherwise, all the baggage of the past surfaces every time a round, extension, or exit comes up.

Getting there means figuring out who the insiders are that still have capital to support the company, whose signatures are required to authorise any restructuring, and who needs to receive something in exchange for cooperation. It’s messy work. But the alternative is a company that keeps getting blocked by its own history, deal after deal, even when the fundamentals are strong.

GLP-1 is the AI of food, for better and worse

When Steven said “GLP-1 is our AI,” he meant it both ways. On one hand, GLP-1 drugs represent a genuine platform shift. The percentage of the population on these drugs is rising fast, and the behavioural changes, smaller portions, reduced appetite for ultra-processed food, and a shift toward protein-dense options are already showing up in purchasing data.

Importantly, the effect radiates beyond the individual user. As households adjust, the ripple is larger than the headline number suggests. A lot of Siddhi’s CPG portfolio, protein-forward, better-for-you brands, are well-positioned. And as consumer purchasing shifts, ingredient adoption on the B2B side will follow faster.

But on the other hand, just as every company in the AI era is suddenly an “AI company,” every food startup now has always been a “GLP-1 companion.” The positioning is often reverse-engineered after the fact. There’s real bubble risk here, and Steven has seen the pattern before.

On AI itself, he understands the productivity case for AI-powered CPG salesforce tools, one human managing an army of agents doing prospect research, initial outreach, and pipeline filtering. What he doesn’t see is the investment case. “I struggle to see the investment case in that software company,” he said. “I don’t see a moat.” These tools are, in his view, “next-generation services businesses” that will live or die on sales execution rather than on proprietary technology.

And the platform risk is real. Because the underlying models are controlled by a few large players, he expects those LLM providers to watch which use cases take off and then “eat the lunch of their wrappers” by adding the most successful functions natively, in the same way Shopify absorbed some of its own app ecosystem into the core product.

That does not mean AI is irrelevant to food tech. Where he sees the genuine case is in tools that remove hardware from the loop in bioprocess development: running virtual bioreactor experiments in silico in seconds instead of in steel over weeks, with outputs reliable enough to drive real decisions.

For a cash-strapped sector, anything that compresses the iteration cycle and reduces the amount of physical experimentation needed changes the shape of budgets and timelines in fermentation, cell culture, and other process-heavy categories. Those tools are harder to commoditise because they are built on the actual physics and biology of the problem, not on a prompt. That is the difference between AI that creates a moat and AI that just borrows one.

If you enjoyed this article, you might also like

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Share

In Issue #140 of Better Bioeconomy, I want to talk about a term that I have been coming across frequently lately: Physical AI.

Tractor companies, laser weeding startups, and drone platforms are all using it. At CES 2025, NVIDIA’s Jensen Huang described Physical AI as “the next frontier of AI,” and the phrase spread from there with the momentum that only a newly mainstream tech label can generate.

Every few years, the vocabulary in this agrifood tech space refreshes. From “precision agriculture” to “smart farming” to “ag robotics” and now “physical AI.” Some of those rebrands tracked real changes in what was happening on the ground, but some were mostly new dressing on a familiar pitch.

So which is this one? The answer, after reading into the underlying technology and watching which companies are doing what: partly both. Some of it is marketing. But underneath the label is a real architectural shift in how robotic intelligence is built and deployed. Recognising that shift helps me understand which developments in the space deserve attention right now.

Let’s dig in!

The previous wave of ag robots shared a structural problem that engineering alone couldn’t fix

The first thing to understand about Physical AI in agriculture is that it comes after a decade of promises that mostly didn’t deliver. “Wave 2” ag robotics, during 2015-2022, was built around a plausible-sounding bet: attach a computer vision model to a robot arm, and it can pick fruit, identify weeds, and scout for disease. Many companies launched on this premise. Several raised significant capital. Most either failed outright or remain stuck searching for commercial scale.

The failures get attributed to various causes in post-mortems: difficult outdoor conditions, high hardware costs, farmer scepticism, pandemic disruption, and poor funding timing. Hardware economics were a real constraint. So was the fundamental difficulty of dexterous manipulation in unstructured outdoor environments, a mechanical challenge that still needs work in applications like harvesting today.

These all played a role. But reading across the category, another constraint shows up repeatedly alongside them: the intelligence itself was often narrow and hard to transfer.

A strawberry-harvesting robot trained in one greenhouse did not necessarily transfer well to a different layout, variety, or growing region. An apple-picking robot optimised for one orchard geometry couldn’t adapt to a new farm configuration without restarting the data collection and model training process. Every new application was, in effect, a cold start.

If a system has to be heavily retrained or reconfigured every time the crop, geography, or workflow changes, its commercial scope narrows quickly. Layer on seasonal data collection constraints (you get one shot per year at harvesting strawberries), hardware costs that haven’t yet benefited from scale, and maintenance complexity in outdoor conditions, and the commercial window becomes much tighter than early narratives suggested.

This is not to say that transferability was the only thing that broke the previous wave. It was that narrow, non-transferable intelligence that appears to have been one of the recurring structural constraints that made all the other problems harder to overcome.

Physical AI stacks three technologies, and the VLA layer is the architectural shift

As mentioned at the start, the term “Physical AI” was popularised by NVIDIA’s Jensen Huang at CES 2025, where it served as the organising concept for NVIDIA’s push into robotics hardware: the Isaac platform, the Jetson edge computing chips, and the model called Cosmos.

Physical AI, as NVIDIA defines it, means AI that doesn’t just process information but perceives the physical world through sensors and acts in it through motors and actuators. The distinction from software AI is the closed loop between perception and physical action: a system that reads the scene and then moves, rather than reading the scene and producing text or a label.

That loop is built from three layers of technology that have been developing in parallel.

1. Foundation model

Foundation models are large AI systems pre-trained on enormous datasets, general enough to be adapted to a wide range of tasks without retraining from scratch. GPT, Gemini, and LLaMA are foundation models for language and multimodal understanding.

The core property is transferability. A foundation model trained on broad data can be fine-tuned on a small amount of task-specific data and perform well, instead of needing a full training run for each new application.

This is the architectural property that reduced the cold start problem in natural language processing around 2020. In robotics, the equivalent shift is underway. The cold start problem is not yet eliminated, but it is being structurally reduced.

2. Vision-Language-Action (VLA) model

A VLA takes three inputs: a camera feed showing what the robot sees, a natural-language instruction describing what to do, and the robot’s current sensor state. It produces one output: the specific physical action to take next, expressed as motor commands, joint angles, or gripper positions.

VLAs extend foundation models into the physical world by adding that action output layer. A robot controlled by a VLA doesn’t require hand-coded rules for every situation. It reads the scene, parses the instruction, and reasons about how to move.

A VLA-controlled robot is like an experienced farmhand who has worked across dozens of different properties and crops and can read a new field after a short briefing, versus one who only ever trained on a single farm and needs months of retraining when anything changes.

The key property is that VLAs generalise across tasks. Physical Intelligence’s π0 model, not an agricultural application but a clear demonstration of what this architecture can do, was trained on data from seven different robot configurations across 68 tasks. The intelligence it encodes can generalise across hardware types and be fine-tuned to new manipulation tasks with far less task-specific data than training from scratch.

3. World foundation model

A big, persistent obstacle in training physical robots is that real-world data is expensive and slow to collect. In agriculture, it is also seasonal: you cannot train a strawberry-picking robot in January.

Cosmos, trained on 20 million hours of real-world video comprising 9,000 trillion tokens, can generate photorealistic synthetic environments for robot training. It is like a flight simulator for farming: instead of waiting for the season to collect real training data, you build the field in software and run thousands of synthetic harvests.

The goal is to compress the gap between simulation and reality enough that robots trained in virtual environments transfer reliably to the field. Danny Bernstein, CEO of The Reservoir, an agrifood robotics investor and startup incubator, described one operational signal I found very interesting: “sim-first development” is allowing teams to “iterate in production, not just in labs, and that changes what’s possible in 2026.”

Companies building on synthetic training environments are not constrained to one training cycle per year, which has historically been one of agriculture’s hardest structural limits on model development speed.

Subscribe now

The question that separates real Physical AI from the rebrand is whether the intelligence can transfer

Before getting into specific companies, it helps to be clear about what Physical AI actually covers in agriculture, because the label is doing a lot of work across three different categories.

The first is manipulation robotics. These are systems that physically interact with crops through picking, laser weeding, or precision intervention at the plant level. This is the category where VLA models are most directly relevant.

The second is autonomous field vehicles. Tractors and spray platforms that navigate autonomously but don’t perform dexterous manipulation. Their AI architecture is closer to autonomous vehicle technology, combining GPS/RTK positioning, obstacle detection, and path planning.

The third is AI-guided precision application. Computer vision systems that identify a target (weeds, diseases, nutrient deficiencies) and trigger a response.

All three carry the Physical AI label. The cold-start argument applies most acutely to manipulation robotics, and commercially deployed VLA-based systems in agriculture are still early.

The test for whether a company is participating in the architectural shift or borrowing the vocabulary is: are they building intelligence that generalises across crops, contexts, and geographies, or building task-specific models that require restarting from scratch for each new application? The former inherits the generalisability that foundation models provide. The latter inherits the cold-start constraints that plagued Wave 2 ag robotics.

By that test, Carbon Robotics’ announcement in February is one of the more instructive examples of agricultural Physical AI to date. The company launched what it calls the Large Plant Model (LPM), which it describes as a foundation model trained on 150 million labelled plant images across crops, weeds, soil types, climates, and growth stages worldwide.

Rather than training a separate model for each new crop context, the LPM provides a base that Carbon says any new field configuration can be added to using two to three reference images, in minutes rather than weeks of data collection. Carbon describes it as “the foundation for Carbon AI, the decision-making brain operating across all of Carbon Robotics products.”

The cold start problem it is claiming to have resolved is specifically at the plant recognition layer: the system generalises across plant types rather than being retrained per crop. That is a meaningful shift from Wave 2 architecture.

The companies passing the architecture test are still a limited subset of the full field now using “Physical AI” in their positioning. Most agricultural robots in commercial deployment today were built on Wave 2 architecture: task-specific models, one crop, one context.

Commercially deployed VLA-based systems in agriculture don’t yet exist at a meaningful scale. Carbon’s LPM is one of the most credible examples of foundation-model thinking applied to an agricultural perception task.

The next proof point will likely come from manipulation robotics, and from harvesting specifically, because that is the category where task-specific architecture struggled and where the transfer properties of foundation models change the economic math most significantly.

I think there is a reasonable case that agriculture may be structurally harder than the domains where foundation models have delivered results so far. Outdoor growing environments are much more variable than warehouses or factory floors. The sim-to-real gap, the difference between what a model learns in simulation and how it performs in an actual field, remains a challenge.

Seasonal data constraints mean that even a well-designed foundation model may need several growing cycles to accumulate enough real-world fine-tuning data to be reliable. And the mechanical challenges of dexterous manipulation in unstructured environments, soft fruit, variable canopy, unpredictable terrain, have not gone away just because the perception layer has improved.

The categories reaching commercial scale show where Physical AI gains traction next

The investment data from 2024 and 2025 tells a useful story about which categories of agricultural robotics have reached commercial traction, and what they have in common.

Weeding and AI-targeted spraying lead the field. Carbon Robotics raised a $70 million Series D in October 2024, with more than 100 LaserWeeders deployed in customers’ hands. Ecorobotix, the Swiss AI-targeted micro-spraying company, raised $150 million across its Series C and D rounds in 2024 and 2025 and is scaling commercial operations globally. Its systems cut herbicide volumes by up to 95% through AI-targeted application rather than blanket coverage. John Deere’s See & Spray system offers the most instructive scale data in the category: by 2025, it was operating across more than five million acres, achieving herbicide reductions of close to 50% on average according to Deere’s own figures.

What these categories share is a favourable task structure. The decision the robot makes is narrow and often binary at the plant level: weed or crop, spray or skip. The outcome is directly measurable in input savings. And the task is repeatable across field conditions without requiring the kind of open-ended dexterity. That combination of narrow decision, measurable outcome, and repeatable task is the recipe that has translated agricultural robotics into verified commercial ROI.

Harvesting has not yet followed the same trajectory, and I think the reason is structural. It combines perception under variable field conditions, delicate manipulation, throughput requirements, seasonal data bottlenecks, and high reliability expectations in a labour-intensive workflow.

Three key takeaways

1. Transfer architecture is the filter that separates the shift from the rebrand

“Physical AI” will shortly be on the deck slide of every ag robot company with a camera and a motor. The terminology is not a useful filter. The question that cuts through is architectural: what happens when you move this system to a new variety in a new geography?

If the answer involves months of data collection and retraining, that is Wave 2 architecture with a new label. If the system can be fine-tuned on a small dataset and deployed in days, that is a fundamentally different scaling proposition with different economics.

2. Weeding’s commercial traction points to where the next category breaks through

Maarten Goossens, Founding Partner at Anterra Capital, identified the critical transition in a recent AgFunder investor survey: AI is moving “from models to applications embedded in workflows.” Agrifood, he argues, is “one of the largest real economy opportunity sets where workflows are still under-digitised, so ROI is unusually tangible.”

That tangibility is what made weeding and precision spraying the first categories to reach commercial scale, and it offers a blueprint for founders working in adjacent areas.

The categories closest to the same kind of breakout share specific structural properties: a narrow, well-defined decision at the plant or field level, an outcome that can be measured directly in input cost or yield terms, and a task that repeats across growing conditions. Harvest timing assessment, disease identification, and targeted nutrient application each have paths to that structure.

3. The data flywheel is the asset that compounds, and the robot is the collection mechanism

Carbon Robotics’ Large Plant Model, trained on 150 million plant images and growing continuously through its deployed fleet, is a useful frame for thinking about where durable competitive advantage sits. Deployed systems collect labelled field data, which improves the model, and better models make future deployments faster and more capable.

This flywheel logic is more applicable to agrifood because agricultural data is hard to get, seasonal in nature, and highly contextual. Companies that own large, clean, well-labelled agricultural datasets hold the fine-tuning substrate that new entrants will struggle to replicate from scratch.

The underlying constraint is real: agricultural data remains sparse, fragmented, and poorly digitised across most of the sector. AI is only as good as what it is trained on, and a model built on incomplete or noisy data compounds those limitations at scale. The durable competitive question for this category is which company owns the data that makes the next generation of models work.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it!

Share

Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Disclaimer: The views and opinions expressed in this newsletter are my own and do not necessarily reflect those of my employer, affiliates, or any organisations I am associated with.

Welcome to Issue #139 of Better Bioeconomy, weekly insights on how tech is reshaping food and agriculture for better human and planetary health. Thanks for being here!

The Strait of Hormuz blockade, in place since March 2, has removed 43% of global seaborne urea from the market. Urea prices have jumped over 30% since. Northern Hemisphere planting decisions are weeks away.

Against that backdrop, this issue carries three different approaches to reducing dependence on conventional nitrogen supply chains. PlasmaLeap’s A$30 million raise produces nitrogen from air, water, and renewable electricity in modular on-farm units.

Crop Diagnostix, which launched commercially this month, takes a different angle: RNA sequencing of plant leaf samples to tell farmers precisely when they don’t need to apply nitrogen. In field trials that translated to a 38% reduction in nitrogen use alongside an 8% yield increase.

Vietnam’s low-emission rice programme shows what systematic reduction looks like at scale: 354,000 hectares reached two years ahead of target, a 38% reduction in nitrogen application per hectare, a 13.4% rise in farmer income.

The pattern also showed up outside this issue: Birufinery received investment this week for seaweed-microbe biostimulants claiming 30-50% fertiliser reduction in tropical plantation crops, and Rovensa Next launched a dual-microbial inoculant in Brazil targeting nitrogen fixation and phosphorus solubilisation across ten crops.

None of these is a complete answer to a supply crisis. PlasmaLeap is commissioning first-of-a-kind systems. Crop Diagnostix is launching in corn and potato only. Vietnam’s results came with guaranteed offtake that most markets won’t replicate. But the blockade has made the commercial case more convincing and urgent for all of them.

BIO BUZZ

Products, partnerships, and policies

🏭 Latam sugar giant Magdalena is building a 650,000-L precision fermentation facility in Guatemala, targeting high-value molecules

• Founded in 1983 and still family-owned, Magdalena has long monetised sugarcane byproducts through molasses fermentation and bagasse-fired energy. Its new biochem unit goes further: a 650,000-L CMO facility breaking ground in May 2026, starting at 50,000 L with a 2,000-L pilot plant and scaling to full capacity by 2029.

• The competitive case rests on three structural advantages: captive feedstock (sugar), existing bioprocessing infrastructure, and the acquisition of Biorbis, an R&D lab in Portugal previously owned by Amyris. Downstream kit will cover centrifugation, microfiltration, and spray drying, enabling industrial enzymes and dry protein formats.

• In addition to the CMO play, Magdalena has launched Proteva, upgrading spent distillery yeast into functional feed and pet food ingredients via enzymatic hydrolysis. The longer-term ambition is to own proprietary molecules, focused on nature-identical compounds that are hard to scale conventionally.

Source: AgFunder News

Thoughts 🤔

Four decades of sugar cash flows, not venture capital, are funding this facility. That removes the liquidation timeline and growth mandate that push most fermentation startups toward premature scale. The feedstock is already integrated into core operations, so the marginal cost of fermentation inputs looks structurally different from a standalone build.

The CMO-first sequencing adds to this discipline. Contract manufacturing revenue from clients like Coca-Cola and Bacardi validates the process stack before Magdalena commits capital to proprietary molecule development. The Biorbis acquisition is worth noting here: Amyris’s Portugal R&D lab, picked up during bankruptcy in 2023, likely gives Magdalena strain development capability for that longer-term play. Amyris failed partly because it tried to own strain IP, manufacturing, and consumer brands simultaneously without the cash flows to sustain all three. Magdalena is taking the science without the overhead.

I think what determines whether the proprietary pipeline moves from ambition to funded R&D is how CMO margins hold up. If contract manufacturing stays profitable long enough to self-fund the next phase, the sequencing makes sense. If margins compress as more fermentation capacity comes online, Magdalena faces the same capital problem it was designed to avoid.

🍫 Puratos to launch the world’s first cell-based chocolate for commercial use by end of 2026

• Belgian confectioner Puratos, which invested in California Cultured through its VC arm Sparkalis, will now bring a cell-based cocoa product to its US customer base by the end of the year.

• Puratos frames cultured cocoa as a “climate-independent and sustainable complement” to conventional farming, citing more consistent quality and supply. It is drawing on its flavour development and sensory science expertise to ensure the ingredient meets food manufacturer expectations.

• For chocolate makers, the appeal is a cocoa ingredient free from supply volatility. The commercial launch moves cultured cocoa from scientific proof to a dependable ingredient that manufacturers can plan around.

Source: Green Queen

Thoughts 🤔

Two major chocolate incumbents anchoring commercial products around the same startup, before FDA GRAS clearance is even secured, is interesting. Puratos is both investor via Sparkalis and first commercial channel for the US market. Meiji, Japan’s largest chocolate company, signed a 10-year supply agreement targeting both the US and Japanese markets. Together, that is a multi-anchor launch strategy that distributes commercial proof across geographies and buyer types before the product is fully approved.

That structure matters for plant cell culture broadly. The technology has been validated for niche cosmetic and supplement ingredients, but a high-volume food commodity like cocoa is a harder test. Flavour complexity, performance specs across bakery and confectionery, and procurement scrutiny from professional food manufacturers all set a higher bar.

If both launches perform on taste and batch consistency, California Cultured becomes the one of clearest proof-of-concept for plant cell culture moving from niche botanicals into commodity food ingredients. If either struggles at scale, the window narrows considerably, and the sector’s broader ambitions take longer to prove out.

🔁 California Cultured completes first commercial-scale run of plant cell-cultured cocoa using $3k reusable plastic bioreactors

• Plant cell culture has historically relied on expensive stainless steel tanks. California Cultured’s proprietary rigid plastic bioreactors run at 2,000-L, can be reused thousands of times and cost ~$3k each. A comparable stainless steel tank could cost $500k to $1M.

• Automated control software lets large numbers of reactors run with minimal labour, compressing both capital and operating costs. The scale-out model deploys many inexpensive modular units in retrofitted light industrial space, avoiding the centralised mega-plant economics that have made biomanufacturing difficult.

• The first product, a cultured cocoa powder with zero lead and cadmium, is positioned as a flavanol-boosting health ingredient rather than a direct cocoa replacement.

Source: AgFunder News

Thoughts 🤔

Most biomanufacturing scales by building fewer, larger, more expensive units. California Cultured runs in the opposite direction: many small, cheap reactors in retrofitted light industrial space, with automated control software compressing the labour cost that usually scales with unit count. That is a scale-out model, and it appears in other technology cycles at the point where centralised infrastructure proves too capital-intensive to deploy quickly enough.

The CAPEX compression matters beyond cocoa. Stainless steel infrastructure costs have kept commodity food applications using plant cell culture structurally out of reach. If $3k reusable reactors hold at 2,000L and above, the cost game for plant cell culture across a wider range of crop inputs changes significantly.

But whether the architecture generalises is the open question. Reactors optimised for a single cocoa cell line are a specific case. Different plant species, growth parameters, and contamination profiles may each limit transferability. If those constraints are manageable, California Cultured may have built a manufacturing template as much as a product.

🛒 Nestlé launches Snack Vibes in Germany using Planet A Foods’ cocoa-free chocolate alternative

• Snack Vibes goes on sale next month under Nestlé’s Choco Crossies brand in Germany, targeting Gen Z in three flavours: classic, hazelnut, and salted popcorn caramel. Each 100g pack retails at €2.79 and is made with Planet A Foods’ ChoViva instead of conventional chocolate.

• ChoViva is produced by fermenting and roasting sunflower and grape seeds to replicate cocoa’s aroma, flavour, and texture, combined with plant-based fats and sugar. It functions as a 1:1 cocoa replacement in vegan, dark, milk, and white chocolate formats.

• Planet A Foods has expanded its Pilsen production facility from 2,000 to 15,000 tonnes annually, and has entered into a commercial partnership with Barry Callebaut. ChoViva is already in over 120 products across 10 countries and more than 100,000 stores globally, with partners including Lindt, Aldi, Lidl, Rewe, and Kaufland.

Source: Green Queen

🐶 Magic Valley launches Australia’s first cultivated meat for dogs

• Training Bites blend cultivated pork with oat flour, apples, sweet potatoes, and brewer’s yeast. Sold in limited drops at A$14.95 per 50g pack, they are the first cultivated meat product available for pets in Australia and Magic Valley’s first commercial proof point for its broader pork platform.

• Pet food regulation in Australia sits under voluntary industry standards rather than government approval, giving the company a faster path to market than human cultivated meat requires. Magic Valley is using Rogue Pet to build real operating capability across production, fulfilment, and repeat purchase.

• The first drop selling out in a week gives Magic Valley something rare in cultivated meat: real customer demand and early revenue. Future drops, expanded market reach, and B2B supply to other pet food brands are on the roadmap.

Source: Green Queen

🌾 Vietnam has nearly doubled its low‑emission rice target two years early by reducing adoption risk

• The programme set out to cover 180,000 hectares of low-emission rice between 2024 and 2025. By the end of 2025, it had reached 354,000. The acceleration largely came from a single structural change: guaranteed offtake for every tonne produced, removing the market risk that typically stalls smallholder adoption.

• Participating farmers adopted mechanised seeding and improved water management, cutting emissions by 3-4 tonnes of CO₂e per hectare against conventional growing. Average farmer income rose 13.4%. Vietnam also piloted its Measurement, Reporting and Verification (MRV) process across 11 model areas, with both staff and farmers able to run it independently.

• The government plans to add 800,000 more hectares before 2030. Over 5,000 hectares are already certified under a "green rice" label, positioning the crop for market differentiation and climate-aligned trade premiums.

Source: AgTechNavigator

Thoughts 🤔

Smallholder adoption of known-good farming practices almost always stalls at the same point: who absorbs the market risk when the practice works but the price doesn’t.

What Vietnam assembled was a risk transfer architecture rather than a technology programme. Guaranteed offtake removed farmer downside. World Bank carbon payments added a revenue stream not contingent on commodity prices. A green export premium, with certified rice reaching $820 per tonne against $400-450 for conventional Vietnamese fragrant rice, layered a third incentive on top. The farming practices were already well-documented. The bottleneck was never agronomic.

That distinction is important if we want to replicate this across regions. The constraint being solved here is an institutional one. Other rice-growing countries in Southeast Asia face the same adoption stall on similar known practices. Can the multilateral financing and offtake commitments that made Vietnam’s architecture work can be assembled elsewhere without compressing the premium?

🌽 Crop Diagnostix uses RNA sequencing to flag crop disease up to four weeks early

• The California-based startup borrowed a diagnostic approach from human medicine and applied it to crops. It reads which genes are active in plant tissue to detect nutrient deficiency, disease pressure, drought, and water stress before any visual sign appears. By the time symptoms show, the damage is already done.

• Growers send weekly leaf-punch samples in vials and receive results within 48 hours across four indexes: yield response, nutrient status, pathogen stress, and water stress, trained on thousands of real-world RNA samples from corn and potato fields. In field trials, the system detected disease four weeks before visible symptoms.

• Pilots in corn and potato showed an 8% yield increase and 23% reduction in inputs, including a 38% cut in nitrogen, translating to a claimed $720/acre net profit advantage. At $500-700 per field per season, the company is launching commercially in 2026 through agronomy partners including CHS and Wilbur-Ellis.

Source: AgFunder News

Thoughts 🤔

Current crop agronomy is typically organised around visible symptoms. You scout, you see something wrong, you respond. RNA diagnostics flip that sequencing: reading gene expression before any damage appears means the decision window shifts from damage control to prevention.

The four-week lead time is what makes this practically meaningful. Most fungicide and nitrogen timing decisions are already constrained by application windows and crop growth stages. A four-week signal opens up interventions that simply aren’t available when diagnosis happens at symptom onset.

The harder question, I think, is behavioural. Pre-symptomatic data asks growers and agronomists to act on an invisible signal. That’s a different decision-making posture, and the distribution channel here (retailer-affiliated agronomists at CHS and Wilbur-Ellis) inherits grower trust but still has to translate an RNA report into a management change.

If the platform performs consistently beyond the corn and potato pilots, the adoption constraint is likely less about the science and more about whether agronomists are trained and incentivised to use it that way.

BIO BUCKS

Funding, M&As, and grants

🍌 Tropic raised $105M Series C to scale gene-edited banana varieties and deploy TR4-resistant Cavendish from 2027

• The UK-based company has regulatory approval for non-browning and extended shelf-life bananas in five markets. The non-browning variety knocks out the oxidation enzyme, unlocking cut fruit applications. The shelf-life product suppresses ethylene production, adding 10+ days to supply chain windows and cutting cold transport costs.

• TR4 resistance is built on Tropic’s patented Gene Editing Induced Gene Silencing (GEiGS) platform, which edits non-coding genes. Instead of searching through 30,000+ candidate genes for a knockout target, GEiGS redirects the banana’s own RNAi machinery to attack a specific gene inside the fusarium fungus. The approach is tissue-specific, avoids GMO classification, and doesn’t require a full gene knockout.

• TR4 threatens a $25bn industry across Southeast Asia, Africa, and Latin America. Tropic is establishing a mother plantation ahead of commercial deployment from 2027, with capital also funding Black Sigatoka resistance, rice, and new crops. GEiGS is already licensed to Corteva for corn and soybean, British Sugar for sugar beet, and Genus for livestock disease resistance, pointing to platform value well beyond bananas.

Investors: Forbion Bioeconomy Fund, Corteva, IQ Capital, Just Climate, ABN Amro, Invest International, Temasek, and more.

Source: AgFunder News

Thoughts 🤔

The banana supply chain cannot switch away from Cavendish at the pace TR4 is spreading. Variety standards, handling protocols, packaging, and retail contracts are all tuned to a single cultivar, and switching is not realistic even under an existential disease threat. That structural lock-in is what makes TR4 resistance a sale into a captive market. Whoever can fix Cavendish in situ is filling a gap with no substitute.

The oversubscribed round in a compressed funding environment makes more sense in that context, though the cap table suggests more than one thesis at work. It spans a commodity sugar trader (Sucden), development finance institutions (Invest International, ABN Amro), a climate fund (Just Climate), and a strategic ag company (Corteva), capital types that rarely share a syndicate. That breadth points to food security, supply chain resilience, and platform licensing all being priced in simultaneously.

GEiGS is also being deployed well beyond bananas. The platform is already licensed to Corteva for corn and soy, British Sugar for sugar beet, Genus for livestock genetics, and Syngenta Vegetable Seeds, four distinct applications across plant and animal species, before the banana product has reached commercial scale. The platform is accumulating proof of concept across a wider crop base than any banana franchise alone could justify.

But what determines whether GEiGS becomes primarily a licensing platform or a crop company with a licensing arm is whether the banana deployments generate the field-scale performance data that licensees in higher-volume crops need to double down on their commitments.

⚡ PlasmaLeap raised A$30M Series A to deploy zero-emissions, on-farm nitrogen fertiliser production from air, water, and renewable electricity

• The University of Sydney spinout uses a patented plasma reactor to convert air, water, and renewable electricity into ammonia and nitric acid with no fossil fuel inputs. Its modular systems are designed to slot into existing fertiliser supply chains and can be deployed directly on farms or at local hubs.

• Conventional nitrogen production relies on Haber-Bosch and long supply chains. Nitrogen fertiliser accounts for ~2.5% of global CO2e. In parts of sub-Saharan Africa, retail prices reach nearly double the world free-on-board price because of importation, logistics, and financing costs. Decentralised production addresses both problems at once.

• Capital will fund first-of-a-kind hubs in New South Wales and Tasmania. The company expects carbon credits from its decarbonisation impact, with the platform producing synthetic hydrocarbons from biogas or syngas in the longer term.

Investors: Gates Foundation, Investible, Yara Growth Ventures, Twynam, GrainCorp Ventures, and more.

Source: PR Newswire

Thoughts 🤔

Nitrogen fertiliser in parts of sub-Saharan Africa retails at nearly double the world free-on-board price. That gap is not a production problem, it is logistics, importation financing, and supply chain intermediation. Decentralized plasma production does not need to beat Haber-Bosch on synthesis cost to win in those markets. It only needs to undercut the landed cost of imported product, which is a lower bar than the green ammonia narrative typically frames it.

That distinction could reshape the commercialization sequence. In markets where nitrogen deficit directly constrains yields, and where the premium over production cost is absorbed entirely by the supply chain rather than growers, proximity to the farm is itself the value proposition. The technology does not need carbon credit revenue or a green premium to be viable.

The Gates Foundation anchoring this round alongside Yara and GrainCorp suggests the investor syndicate is pricing in both development market deployment and existing agricultural supply chain integration simultaneously. Those are different go-to-market paths with different procurement rails, timelines, and success conditions.

🥛 Verley raised €32M Series A to scale its precision-fermented whey proteins and launch in the US by year-end

• The French startup’s near-term commercial focus is on protein shots, where its purified beta-lactoglobulin delivers high protein concentration while keeping clarity, solubility, and stability intact. First products featuring FermWhey are expected to hit the US market by the end of 2026.

• Verley is operating as a B2B ingredient supplier, targeting food manufacturers and brands that can apply their own formulation expertise and market infrastructure. To scale, it is using a hybrid model combining its own pilot facility with industrial manufacturing partners in Europe and North America, before transitioning to large-scale facilities capable of multi-thousand-tonne annual volumes.

• Following its US launch, the company is advancing regulatory pathways in Europe and the Middle East. The oversubscribed round was won on execution: FDA clearance, validated industrial performance, a growing IP portfolio, and commercial demand that already exceeds current production capacity.

Investors: Alven, Blast, Bpifrance, Sofinnova, Sparkfood, Captech, and Founders Future.

Source: Green Queen

🍼 Guoke Xinglian raised $14.6M Series A+ to scale up precision-fermented human lactoferrin with human-identical glycosylation

• Lactoferrin is the most abundant immune glycoprotein in breast milk, revered for gut health and iron transport, but present in cow’s milk at concentrations 10x lower than in human milk. Beijing’s Guoke Xinglian uses precision fermentation to produce a bioidentical version, more effectively absorbed and gentler than its bovine counterpart.

• The key challenge is glycosylation, attaching sugar molecules to the right positions on the protein, which affects its folding, stability, antigenicity, and resistance to breakdown. The company claims to be the only firm to express human milk glycoproteins via food-grade yeast, using a Kluyveromyces martensii chassis with glycosylation editing.

• Human lactoferrin has entered trial production, with Guoke Xinglian targeting both domestic and overseas B2B supply to infant formula, women’s wellness, health supplement, and functional nutrition brands. The fresh capital will fund AI-driven platform scale-up and global market expansion.

Investors: Shanghai Zhuiguang Julian, Lenovo Capital, Changjin Holdings

Source: Green Queen

🎯 AgZen raised $10M Series B to scale real-time spray monitoring that cuts chemical inputs by up to 50% without yield loss

• The MIT spinout’s RealCoverage system bolts onto any sprayer and uses high-speed cameras to track droplet coverage across the leaf at speeds up to 18mph, detecting droplets as small as 150 microns. It shifts spray management from gallons-per-acre rules to micro-ounces per square millimetre of leaf surface.

• The AI layer processes live coverage data against variables including nozzle type, pressure, boom height, canopy structure, and wind speed, then issues specific real-time adjustments to the operator. A second product, EnhanceCoverage, due commercially in 2027, wraps droplets in adjuvants at the nozzle rather than in the tank, enabling further input reductions.

• The system covered one million acres in 2025 and has commitments for over two million acres across the US, Australia, and Argentina in 2026. The company also announced a strategic partnership and investment from Syngenta Group Ventures, signalling that agchem sees real-time droplet intelligence as infrastructure.

Investors: DCVC Bio, Material Impact, Astanor, and Syngenta Group Ventures.

Source: AgFunder News

Thoughts 🤔

Spray management in commercial agriculture has been governed by gallons per acre for decades. RealCoverage shifts the unit to micro-ounces per square millimetre of leaf surface, and that difference matters for more than farm efficiency.

When coverage is measurable, what growers are buying changes. Volume procurement becomes performance procurement. That creates conditions for premium pricing on high-efficacy formulations. The competitive axis shifts from price per litre to ROI per leaf.

The agchem response to this is visible in the cap table. Syngenta sees RealCoverage as infrastructure for how crop protection products are “tested, sold, and eventually formulated.” What that means in practice is a product development dataset that does not otherwise exist: 150-micron droplet coverage mapped against canopy structure, product, and outcome across millions of acres. Corteva also signed a partnership in October 2025. Two competing majors on the same platform is a strong infrastructure signal.

🧫 baCta raised €7M seed to produce astaxanthin via AI-optimised yeast fermentation for cosmetics and nutraceutical markets

• Astaxanthin is a crimson antioxidant used in dietary supplements, cosmetics, and animal feed, but today’s supply is mainly split between petrochemical synthesis and algae cultivation. The Paris-based startup uses generative AI and liquid-handling robotics to rapidly engineer yeast strains that produce astaxanthin in fermentation tanks, iterating continuously on yield and efficiency.

• baCta claims that yeast reaches over 100g/L biomass density in a bioreactor, compared to under 5g/L for algae in photobioreactors. Combined with a proprietary downstream process that avoids supercritical CO2 extraction, baCta’s cost target is in line with synthetic producers. Its astaxanthin is also claimed to be in the free, non-esterified form, making it more bioavailable than the algae version.

• The company targets cosmetics and nutraceuticals, positioning astaxanthin as an underexploited longevity ingredient. Fresh capital will fund scale-up validation and platform expansion to additional high-value ingredients. Regulatory filings are planned through GRAS in the US and the novel food pathway in the EU.

Investors: LocalGlobe, Daphni, and OVNI Capital

Source: AgFunder News

🌱 Amatera raised €6M to accelerate climate-smart crop breeding by screening at the cell culture stage

• Conventional breeding of perennial crops like coffee or wine grapes can take over 20 years and cost millions. Amatera induces spontaneous genetic variation at the cell culture stage using physical and chemical approaches, generating large libraries of non-GMO cell lines that can be sequenced and screened before any plant is grown.

• Standard practice regenerates all cells into plants before genotyping. Amatera inverts this: it isolates individual cell lines, sequences them in culture, and only regenerates the promising candidates into whole plants. That removes the screening bottleneck and cuts time and cost significantly.

• For perennials, the company develops varieties in-house and licenses them. For annual row and vegetable crops, it offers the platform directly to seed companies. Current candidates in nursery trials include a Robustica coffee combining Robusta disease resilience with Arabica taste, a caffeine-free Arabica, and wine grape varieties resistant to downy mildew and black rot.

Investors: Demea Sustainable Investment, Oyster Bay, PINC, Mudcake, Exceptional Ventures

Source: AgFunder News

🧬 Bindbridge raised $3.8M to design molecular glue herbicides that turn weeds’ own protein recycling machinery against them

• The UK-based startup uses molecular glues, which are single small molecules that hijack a cell’s protein degradation system, forcing weeds to destroy their own essential proteins. Unlike PROTACs (Proteolysis-Targeting Chimeras), which are large two-part molecules that are hard to optimise, molecular glues are closer in scale to conventional herbicides and easier to formulate for field conditions.

• The company’s BRIDGE platform uses AI-driven structural modelling to discover and design these molecules, predicting which compounds can bring two target proteins together to trigger degradation. The platform also filters for agronomic properties like molecular size, charge, and hydrophobicity, ensuring designs translate to the field.

• The funding will support co-development partnerships with ag chem companies and the first round of lab testing. Bindbridge’s initial focus is a broad-spectrum glyphosate replacement, with a longer runway into insecticides, fungicides, and sprayable plant traits for drought tolerance and nutrient use efficiency.

Investors: Nucleus Capital, Speedinvest

Source: AgFunder News

Thoughts 🤔

The new herbicide mode of action pipeline is sparse. Bayer’s Icafolin, announced in mid-2025 as the first new herbicide mode of action in over 30 years, won’t launch commercially until 2028 at the earliest. That context matters for Bindbridge because, from my understanding, induced protein degradation would be mechanistically distinct from what is being used commercially, not enzyme inhibition, not membrane disruption, but forcing target proteins to self-destruct via the cell’s own recycling machinery.

The big question is whether the BRIDGE platform can deliver it. Rational molecular glue design has advanced significantly in pharma, with multiple clinical-stage programs and dedicated computational platforms. But that progress is concentrated around well-characterised human E3 ligases with decades of structural data behind them. Plant protein targets don’t yet have that foundation. Whether AI-driven structural modelling can bridge that gap for agricultural biology, where the structural data is thinner, is the assumption the thesis rests on.

But that assumption will be tested relatively quickly. The first co-development partnerships and lab validation data will be the earliest signal either way.

🤖 Living Models raised $7M to build AI foundation models for plant genomics and compress the seed development cycle

• The French startup applies the transformer architecture that powers large language models to plant DNA sequences, treating nucleotides as a language instead of relying on computer vision or phenotype-based analysis. Its BOTANIC series spans models from 100M to 1B parameters trained on 43 phylogenetically diverse plant species, with open-weight models available on Hugging Face.

• Conventional trait discovery requires growing plants and working backwards from phenotypes. In-silico DNA experiments allow researchers to identify stress tolerance and yield stability traits before a plant is grown, targeting a development cycle that currently takes up to 8 years.

• Fresh capital will fund model and platform development. CEO Cyril Véran has framed the plant genomics focus as a deliberate starting point, indicating that BOTANIC is a beachhead for a broader life sciences foundation model strategy.

Investors: Asterion Ventures, The Galion Project, Kima Ventures, STATION F, University of California Berkeley, and more.

Source: iGrow News

🍫 Cellva Ingredients raised $3.8M pre-Series A to scale its cocoa-free chocolate ingredient made from coffee production waste

• The Brazilian startup extracts bioactive compounds from two coffee by-products (parchment and silver skin), microencapsulates them for stability, and supplies the result as a drop-in ingredient replacing up to 70% of cocoa in bakery, confectionery, beverage, and nutrition applications.

• Instead of sourcing virgin agricultural inputs, the startup partners directly with coffee producers in Bahia’s regulated growing regions, building full traceability from farm to finished ingredient. The company’s ingredient, CoffeeCoa, carries a naturally sweet profile and is rich in antioxidants and fibre.

• From gram-scale batches in 2024, the company projects processing around 2,000 tonnes of raw material this year. It is already supplying industrial bakery manufacturers and chocolate processors in Brazil, and is advancing commercial discussions with clients in Europe, Asia, and the US.

Investors: DigiBoard, Air Capital, Rubens Pereira

Source: Green Queen

🌍Join HackSummit 2026, get 15% off your pass using my code

HackSummit returns to Lausanne, Switzerland on April 22-23, 2026 for its 5th anniversary.

Over two days, founders, investors, and operators will explore how industrial capacity and food sovereignty are becoming foundational systems in a climate-constrained world.

Day 1 focuses on unlocking infrastructure, manufacturing, energy, and hard-tech systems that underpin resilience.

Day 2 dives into Food Sovereignty by scaling stable, regenerative, and secure food systems amid climate volatility and supply risk.

Join the builders shaping the next decade of industrial and biological resilience including Altrove, Paebbl, MicroHarvest, Hades Mining, AtomoCooling and Voltrac. As well as Partners and Managing Directors of Visionaries Tomorrow, Voltfang, Space Forge, FORWARD.one, Slate VC, Contrarian Ventures and Unruly Capital.

💰 Use my code 'BetterBio15' for 15% off your pass.

If you found value in this newsletter, consider sharing it with a friend who might benefit! Or, if someone forwarded this to you, consider subscribing.

Share

Subscribe now

Thank you, and have a great day!

Disclaimer: The views and opinions expressed in this newsletter are my own and do not necessarily reflect those of my employer, affiliates, or any organisations I am associated with.

Thanks for being here. In Issue #138 of Better Bioeconomy, I want to share something that has shifted in how I think about failure in agrifood.

For a long time, my working theory when a technology stalled was some version of: wrong market timing, wrong team, wrong capital structure. They were usually a company-level explanation. And often, those explanations are right.

But I kept noticing the same failure pattern across companies with different technologies, different founders, different geographies, and even different capital structures. Company-specific problems explain individual failures. They do not explain why the pattern repeats so consistently across companies that share almost nothing except the system they entered.

The Food and Agriculture Organisation (FAO)’s ‘Transforming food and agriculture through a systems approach’ report gave me the language for what I was noticing. The failure was not always a company-level problem. It was a system-level one: technologies trying to enter systems that were not designed to receive them.

Agrifood systems are not just supply chains

Before the interesting part, some grounding on what we are talking about.

When FAO uses the phrase “agrifood system,” it means something bigger than the farm-to-fork supply chain most people picture. It captures the full journey of food from production through processing, distribution, retail, and consumption, including all the waste that comes out the other end.

But it also includes the people, institutions, incentives, and policies that govern how food moves through that chain, and the relationships between the food system and everything else it touches: health systems, environmental systems, energy systems, water systems, and economic systems.

This interconnectedness is the central fact that makes food system problems so persistent.

Interventions in one part propagate through the system’s connections to every other part, often in ways that were not anticipated. Subsidise fertiliser to boost yields, and you can inadvertently degrade the waterways downstream that farmers depend on. Design food aid to address acute hunger, and you can simultaneously undermine the local markets that would have provided more durable food security. These outcomes do not require bad intentions. They emerge from the structure of the system itself.

This is also why the same intervention can succeed in one context and fail in another with almost identical surface conditions. The difference is rarely the technology. It is which part of the system the intervention enters, and what the downstream effects of that entry are. Most post-mortems on failed agrifood technologies do not ask this question.

Food governance was built to optimise within sectors, not across them

Fragmented food policy is not a coordination failure that could be fixed with better communication between ministries. It is the output of how food system governance was built.

Agricultural ministries optimise for yields and farm incomes. Nutrition departments optimise for dietary outcomes. Environmental regulators optimise for emissions and biodiversity. Trade departments optimise for export earnings.

Each of these actors is doing what their mandate requires, subject to budget cycles, performance metrics, and political accountability structures that reward progress on their specific objective. But they are not rewarded for thinking about what happens when all of these optimisation functions run simultaneously on the same underlying system.

A recent paper in Frontiers in Sustainable Food Systems identifies four overlapping governance challenges that explain why this pattern is so durable. One of them is what the authors call cross-temporal mismatch: the daily demands of keeping food supply moving pull institutions toward short-term decisions, while the structural changes needed to make food systems resilient play out over decades. The actors most responsible for managing the system’s immediate functioning are structurally the least positioned to redesign it.

Compound that with the difficulty of coordinating across distant value chain nodes, the absence of solutions that satisfy stakeholders with compatible priorities, and the complexity of integrating knowledge across communities that rarely talk to each other, and fragmented governance looks less like a political failure than an architectural one.

The result is a consistent pattern of unintended consequences. Productivity programs that increase yields while degrading the soils and water systems that underpin long-term productivity. Nutrition initiatives that ignore the food environments, retailer incentives, and processing economics that determine what people actually eat. Environmental policies that reduce emissions in one part of the system by pushing costs onto smallholder farmers in another.

FAO’s own policy analysis makes the distinction precise. The majority of food policies it surveyed acknowledged that cross-sector involvement was necessary to implement them successfully. Very few had built any institutional structure to carry that involvement out. The acknowledgement and the architecture are two different things, and the gap between them is where unintended consequences accumulate.

I see this dynamic in agrifood startups. A company might have a genuine technology: a biological input that demonstrably improves soil carbon, a digital tool that measurably reduces water use, a novel protein that cuts the emissions intensity of a diet. The science works, and the pilot results are real. But still, the technology stalls somewhere between proof of concept and at-scale adoption because the surrounding system was not designed to absorb it.

Rob Ward’s “Minimum Viable Ecosystem” concept put a useful framing on what I was observing in a guest piece for AgFunder News. He distinguishes between “challengers,” startups that compete against existing systems in isolation, and “disruptors,” who reconfigure the value chain so that all participants benefit from the change. Most agrifood startups launch as challengers and wonder why the system pushes back. The surrounding value chain has no reason to accommodate them and economic reasons to resist.

Most agrifood due diligence asks whether customers will buy the product. But perhaps the more predictive question is whether the value chain surrounding that customer has any incentive to accommodate the technology/product. A challenger can pass every customer discovery test and still stall, because the distributors, retailers, or institutional buyers adjacent to that customer have economic reasons to resist. The failure mode is ecosystem mismatch.

The system rewards the actors and decisions that siloed mandates reward. Technologies that require the system to behave differently face headwinds that often have nothing to do with their technical merit. When one of those technologies eventually stalls, the post-mortem focuses on the company: the capital strategy, the go-to-market, the founding team. But we should also ask if the system was ever designed to receive it.

Coordination architecture determines which technologies get to scale

Rather than focusing on individual technologies, actors, or even sectors, the FAO report focuses on relationships. It identifies ten key relationships within agrifood systems, from production-consumption dynamics to the institutional structures that shape decision-making power, and argues that these relationships are where the real leverage sits. A change at the right relationship node can cascade across the system in ways that a targeted point intervention cannot.

The report then proposes six elements it considers necessary for food system transformation: systems thinking, systems knowledge, systems governance, systems doing, systems investment, and systems learning. All six elements are about how the system is organised: how it generates and shares knowledge, how it coordinates across sectoral mandates, how it designs and funds interventions, and how it adapts when outcomes diverge from plans.

Some of the examples the report draws on illustrate what this looks like in practice. New York City’s public procurement reform simultaneously shifted nutrition outcomes, reduced the climate footprint of the city’s food purchasing, improved equity in food access, and created new market signals for suppliers across a complex supply chain.

Ethiopia developed an integrated national food systems vision that brought together nutrition, climate resilience, food safety, and farmer livelihoods into a single framework, then used it as the organising logic for investment decisions.

Indonesia built economic models to explicitly map the trade-offs between food security, biodiversity, and economic development, creating the knowledge infrastructure needed for cross-system decisions rather than single-sector ones.

A study in Philosophical Transactions of the Royal Society B, led by FAO researchers drawing on food system implementations across Morocco, Costa Rica, Rwanda, and the Pacific Islands, found the same sequencing in each case. Transformation required two things working in tandem: formal governance structures like coordinating bodies and revised policy frameworks, and something harder to fund, the shifts in relationships, trust, and institutional thinking that allow different mandates to align. Neither was sufficient without the other. And in the case examined, the technologies and investment decisions came after that combination existed, not before.

Each created a mechanism that made the system’s trade-offs visible across sectors before decisions were made, and positioned one actor to coordinate above the siloed mandates rather than within them. That sequencing matters more than any of the specific interventions that followed. The technology and investment decisions in each case came after that coordinating mechanism existed. Most agrifood innovation runs the sequence in reverse: the technology arrives first, and waits for the institutional conditions to catch up.

Subscribe now

The forces that built the broken architecture are the same ones that protect it

The system is not failing to change because people lack good ideas about what needs to change. It is failing to change because the same structural features that produced the broken architecture, short political time horizons, fragmented authority, and uncertainty about what will work, are also the features that make redesigning it so difficult. Three overlapping forces explain the bind.

1. Time

Building coordination across institutions, ministries, and sectors takes years. Political cycles, by design, reward visible progress on short timescales, which means the interventions most likely to produce durable food system change are the ones that take the longest to show results.

I see the same mismatch runs through capital markets. Meaningful agricultural innovation may take 10-15 years to mature, against a 7-10 year (sometimes even shorter) VC fund horizon. A fund operating on that clock will rationally favour technologies that can show commercial traction within that window, which means the technologies most likely to produce systems-level change are systematically deprioritised. The capital structure is doing what it was designed to do, but the problem is that it was designed for a different job.

2. Leadership

Systems transformation requires someone, or some institution, with both the authority and the knowledge to coordinate across competing interests and power structures. In practice, those two things rarely sit in the same place.

The actor with political authority to coordinate tends to lack the technical understanding to do it well, the actor with deep technical knowledge rarely has the mandate to impose coordination across ministries or value chain actors. Ethiopia’s food systems vision worked in part because it was deliberately co-convened by two ministries, which gave it both the cross-sector legitimacy and the technical depth that single-ministry leadership could not provide. That kind of deliberate coordination architecture is lacking.

3. Uncertainty

Complex systems do not respond predictably to interventions. A procurement reform that works in New York City may produce different effects in Nairobi. This unpredictability changes what the right institutional response looks like.

Transforming a system under uncertainty requires the capacity to run pilots, embed real-time evaluation, and adjust course before scaling, which is a learning infrastructure that most project cycles and private funding milestone structures are not designed to support.

Three implications that are harder to ignore once you accept the diagnosis

The report’s framework is a policy document, not a startup playbook, so I want to be careful not to overextend its implications. But reading it as someone who spends time with agrifood founders and investors, the direction of the implications feels consistent.

1. Agrifood does not have an innovation shortage, it has an infrastructure shortage for deploying the innovations it already has

The conversation in agrifood tends to centre on what needs to be invented: a more resilient crop, a lower-emissions input, a more efficient supply chain. The FAO framework points in a different direction. The six elements it says food systems need to transform are all about coordination, transparency, and institutional design.

Governance tools that help ministries make decisions across sectoral lines. Knowledge systems that make system-level trade-offs visible before policies are committed. Procurement architecture that closes the gap between a corporation’s sustainability pledge and what its buying team actually does. None of these are biological or digital innovations in the conventional sense, but the framework suggests they may be the infrastructure or ecosystem orchestration tools that determine whether everything else gets to scale.

Agrifood is lacking the integration layer that would allow it to function as a system rather than a collection of point solutions. My understanding of what this implies is still evolving, but the conversations I find most interesting in agrifood right now are not about a new crop or a new molecule. They are about how to change the rules of the game that determine whether any crop or molecule gets adopted at all.

2. One of the biggest risks in agrifood is launching into a system that was not designed to absorb you

The adoption curves for precision agriculture, agricultural biologicals, and climate-smart farming practices all follow a similar pattern. The science often precedes broad adoption by years, sometimes decades. What eventually changes is the incentive structure around it: regulatory clarity, supply chain economics, procurement rules, and subsidy design all shift, and suddenly a technology that was technically proven but commercially stalled finds its moment.

One of the things I find underweighted in agrifood investment narratives is who actually bears the cost of being early. The system benefits when adoption accelerates, but the downside of moving first falls almost entirely on the individual producer: the financial risk, the operational disruption, the reputational exposure if the technology underperforms. Ali Morpeth, co-founder of Planeatry Alliance, put this well in AgFunder News: even producers who understand which direction change needs to go are often paralysed because “the downside risks of moving first remain disproportionately high.”

A farmer’s decision to adopt a new technology is a cash-flow decision made within the tight economics of a single growing season, not an investment decision in the VC sense. A technology that delivers ROI over three seasons requires the farmer to absorb risk for three seasons before knowing if it works. The costs of being early fall entirely on the farmer, the benefits are diffuse and slow to arrive.

Most agrifood due diligence runs in one direction: does the technology work, and will customers buy it. But it needs to run in both. Before asking whether customers will buy, ask whether the value chain surrounding that customer has any economic incentive to accommodate the technology.

A few diagnostic questions to ask: Does regulatory clarity already exist, or does the company have to create it? Do the distributors or institutional buyers adjacent to your target customer benefit from your adoption, or does it threaten their margins? Is there at least one actor in the ecosystem with an explicit mandate to coordinate across the value chain? If the answers to those questions are no, you are not launching into an ecosystem, you are trying to build one while simultaneously selling into it. That is a different company, with a different capital structure and a different timeline, than what most early-stage pitch decks describe.

3. The capital model that built ‘agrifood tech’ is structurally mismatched to the transformation timelines the sector needs

A startup can have product-market fit within a segment while being structurally incompatible with the broader system it is trying to change. A soil health technology can work on farms where incentives reward soil health investment. It will not scale in systems where commodity price dynamics, subsidy structures, and land tenure arrangements mean farmers cannot capture the value it creates.

Part of what has shifted in how I think about agrifood investment is the recognition that the standard VC model is structurally mismatched to the sector’s transformation timelines. As Michael Dean of AgFunder said in a piece for AgTech Navigator: “Going to investors and saying we can deliver 3x returns within 10 years is proving super tough.” The consensus points toward patient capital, government involvement, and strategic corporate partnerships as the funding mix the sector actually needs.

Ward’s Minimum Viable Ecosystem concept is one practical response at the startup level: before asking whether a startup can find customers, ask whether it has identified the smallest viable set of partners, incentives, and institutional conditions that would make adoption rational across the value chain.

This is a problem with a known set of partial solutions. Development finance institutions, blended finance vehicles, and strategic corporate ventures with long horizons exist and are funding meaningful work. Patient capital is not theoretically impossible to construct.

The gap is that the infrastructure for connecting agrifood founders to the right capital structure at the right stage is underdeveloped relative to the volume of companies trying to raise money. Most founders encounter the VC model first because it is the most visible, and discover after two or three rounds that the clock was set for a different kind of problem.

The practical implication: knowing which type of capital is right for the problem you are solving is as important as finding capital generally. A company trying to change a system over fifteen years, funded by investors expecting an exit in six, is in tension with its own strategy. The capital structure should match the transformation timeline, not the other way around.

Food systems were designed for a different job

Food systems were not designed to deliver multiple outcomes simultaneously. They evolved to prioritise one thing above others, usually yield, volume, or cost, and to do so within institutional structures that reward that prioritisation. The expectation that they now deliver food security, climate resilience, nutritional quality, economic fairness, and environmental health at the same time is not unreasonable. But meeting it requires redesigning the system.

The FAO report is one of the more honest assessments I have read of what that redesign requires. The barriers it identifies, misaligned time horizons, leadership gaps, and institutional inertia, are features of the same system producing the outcomes we are trying to change.

If that diagnosis is right, the implication is that the agrifood innovation ecosystem needs to make a larger bet on institutional architecture alongside product development. The field has spent a decade asking what technologies can do. The more important question is what systems need to become before technologies can do it.

The technologies that are currently proven but stalled will not stick around. Without the institutional infrastructure to absorb them, they quietly fade away: the startup runs out of runway, the founding team disperses, the IP gets shelved or absorbed into something adjacent. Founders and investors absorb the cost first, visibly, through failed fundraises and written-off positions. But the quieter cost falls on the farmers and food businesses that never got access to something that worked, and on the next generation of founders who will spend time and capital rediscovering the same dead ends.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Disclaimer: The views and opinions expressed in this newsletter are my own and do not necessarily reflect those of my employer, affiliates, or any organisations I am associated with.

Welcome to Issue #137 of Better Bioeconomy, thanks for being here!

Today, I’m pulling apart one of my favourite annual reads, the Verdant Partners 2025 Global Food & Agribusiness M&A Review & Outlook. The report gives a practitioner’s view that most agrifood M&A commentary lacks. These are people who actually run sell-side and buy-side mandates across seeds, livestock, ingredients, equipment, and everything in between, across 30+ countries.

Reading this year’s edition, one theme kept surfacing across nearly every segment: the biggest M&A story of 2025 was separation. From Corteva splitting into two publicly traded pure-plays, to BASF preparing a Frankfurt IPO for its agricultural division, to Bunge and Viterra finally closing their mega-merger and immediately shedding overlapping assets (as part of antitrust requirements and integration), the industry is undergoing a structural unbundling that I think deserves a closer look.

But here’s what makes this more interesting: at the same time that conglomerates are breaking apart at the top, consolidation is accelerating at the segment level. Cooperatives are absorbing smaller retailers. Feed companies are rolling up regional players. Dairy cooperatives are merging into continent-spanning platforms. The unbundling is happening alongside consolidation, just at a different altitude.

Let’s dig in!

A practitioner’s read on a “suboptimal” year

Before getting into the analysis, a word on the source. Verdant Partners is a food and agribusiness-focused M&A advisory firm with offices across North America, South America, and Europe. They’ve completed over $4 billion in transactions across 30+ countries over 25+ years. (This is not a sponsored post!)

I look forward to their annual review because each section is written by someone with direct deal experience in a specific vertical. Dean Cavey on seeds, Blake Croegaert on grain handling and specialty ingredients, Graeme McCracken on agtech. Each section reads like a candid conversation with someone who just closed a transaction and has opinions about where the market is heading.

That practitioner lens is particularly useful for understanding the 2025 vintage of M&A, because the macro context was, to use Verdant Managing Partner Garrett Stoerger’s framing, “suboptimal.”

For many producers, 2025 felt like a replay of 2024. Commodity crop profitability suffered again from low prices driven by abundant supply and waning demand. High-value markets saw stable growth but felt the squeeze of persistently rising costs. Livestock producers experienced the biggest year-over-year improvement, but only after nearly a decade of pain. The overall industry mood had a carryover effect on everyone, from the farmers themselves to the upstream and downstream businesses that serve them, to the investors looking to deploy capital.

Verdant’s own framing centres on “portfolio rationalisation, capital discipline, and targeted growth.” The unbundling I’m focusing on is one manifestation of that broader rationalisation. But the targeted acquisitions happening simultaneously are the other side of the same coin, and they show up clearly in the data.

The biggest deals of 2025 were about subtraction, not addition

Let’s start with the moves that made the trade press. Corteva announced in October that its board had unanimously approved a plan to split the company into two independent, publicly traded entities.

“New Corteva” will retain the crop protection business, with estimated 2025 net sales of $7.8 billion. “SpinCo” will house the seed business, including the Pioneer brand, with projected 2025 net sales of $9.9 billion. The separation is expected to close in the second half of 2026.

Corteva itself was born from unbundling, spun out of DowDuPont in 2019, specifically to create a pure-play agriculture company. Six years later, the logic is being applied again, one level deeper. The seed and crop protection markets, Corteva’s leadership argues, have evolved to the point where the opportunities ahead for both businesses are diverging.

SpinCo is being positioned as a “classic growth compounder” that could expand beyond corn and soybeans into wheat, cotton, rice, fruits, and vegetables. New Corteva, meanwhile, will focus on operational excellence and biologicals. The underlying message is that these businesses need different capital allocation strategies, different risk profiles, and different investor bases.

Around the same time, BASF announced plans to list its Agricultural Solutions division on the Frankfurt Stock Exchange, with a target of being IPO-ready by 2027. BASF’s ag division generated €9.8 billion in sales in 2024. The company will remain the majority shareholder post-IPO, but the intent is to create a standalone, pure-play agricultural company with its own management board and strategic mandate.

Then there’s the Bunge-Viterra merger, which finally closed in July after years of negotiation and regulatory navigation. The combined entity will generate roughly $100 billion in sales. But the integration hasn’t been purely additive. As part of antitrust requirements and internal strategic review, Bunge divested five Canadian grain assets and its North American dry corn and masa milling business (acquired by GrainCraft).

Three different structures, a spin-off, an IPO, and a post-merger divestiture, all driven by the same underlying logic. The preference for focused, legible businesses is winning out over the conglomerate model at the corporate level.

The mid-market carve-outs that confirm the trend

The headline breakups get attention, but the pattern is just as visible in a series of less-publicised moves across the value chain.

Wilbur-Ellis divested its entire nutrition division to Balmoral Funds LLC, which rebranded the business as Rangen. The deal, which closed in August 2025, included animal feed, pet, and aquaculture ingredients businesses and assets. Wilbur-Ellis had acquired companies like F.L. Emmert (specialty pet ingredients) and Ametza (forage livestock pellets) within the prior two to three years, and both were included in the sale. Whether performance-driven or a strategic pivot back to its core retail business, the move brings more private equity capital into the specialty ingredients sector.

ADM and Alltech announced in September the creation of a new animal nutrition entity with 43 feed mills serving the US and Canada. The joint venture, majority-owned by Alltech, represents ADM stepping back from a segment it has operated in for decades. For ADM, this fits a broader pattern of portfolio simplification. For Alltech, it’s a chance to rapidly scale.

FBN (Farmers Business Network) announced the spin-off of its crop protection sourcing and first-mile logistics subsidiary, Global Crop Solutions. The spin-off creates two distinct companies, separating FBN’s grower-facing platform from its supply chain operations.

And in seeds, both KWS and Vilmorin & Cie (Limagrain) elected to exit the US seed corn and soybean market entirely, selling their jointly held AgReliant Genetics business to GDM, the Argentine plant genetics company. Verdant managed the transaction and describes it as perhaps one of the more surprising events in the seed sector in recent years. Two European seed majors decided that competing in US row crop genetics was no longer core to their strategy, opting instead to concentrate on their European corn businesses and other opportunities.

The common thread across all of these: shed what isn’t core, double down on what is, and make the remaining business attractive to capital that rewards clarity over breadth.

Subscribe now

Conglomerates are breaking apart, but within segments, consolidation is increasing

Here’s where the picture gets more complicated. While conglomerates unbundle at the top, consolidation is intensifying within individual segments, and some of the year’s largest deals reflect this dynamic.

In dairy, Arla Foods and DMK Group announced a merger in April that will create the largest dairy cooperative in Europe, bringing together over 12,000 farmers across seven countries.

In crop inputs, Sumitomo Chemical fully acquired and consolidated its European subsidiaries Philagro (France) and Kenogard (Spain) to build a unified regional hub for its biorational and chemical crop protection business. Verdant’s Luboš Grepl describes this as one of the most significant M&A developments of 2025 in the biostimulants space.

In agriculture retail, CHS completed its $225 million acquisition of West Central Ag Services, further strengthening its geographic footprint and absorbing a smaller cooperative into a larger network.

These moves share similar logic. Arla and DMK are combining to achieve the scale needed to compete in a volatile dairy market. Sumitomo is integrating to build a coherent European platform. CHS is absorbing a smaller player to spread costs and strengthen resilience. At the segment level, scale still creates value, especially in asset-heavy, margin-tight businesses where financial strength determines survival.

At the corporate portfolio level, the trend is toward separation and focus. At the segment level, the trend is toward consolidation and scale.

What changed to make unbundling the dominant strategic response in 2025

So why is this happening now? I see three forces converging.

First, the agricultural economy is bifurcating in a way that punishes sprawl. Verdant’s segment-by-segment analysis paints a stark picture. Commodity grain producers are potentially heading into a third consecutive year of hardship. Meanwhile, specialty segments, including berries, high-value vegetable seeds, premium proteins, biologicals, and healthy snacking, maintain strong margins and healthy demand.

When your portfolio spans both sides of that divide, capital gets misallocated. Profitable divisions subsidise struggling ones. The rational response is to separate, letting each business compete for capital on its own merits.

This tracks with what I explored in Issue #131, where McKinsey’s analysis of 134 publicly traded agrifood companies showed that outperformers move roughly 50% of their capital across businesses over a decade. The Verdant deal log is essentially that thesis playing out in transaction form: the companies that move capital are the ones repositioning for the next cycle.

Second, a new class of buyer has entered the market, and they’re deploying capital with sector-specific conviction. Private equity firms like KKR and Patient Square, sovereign wealth funds like ADQ, and non-agricultural strategics like DENSO are all making investments in food and agriculture. These investors think in terms of focused theses, not conglomerate logic. They want pure exposure to animal protein infrastructure, or vegetable seed genetics, or animal health distribution. Stoerger’s observation that the current environment favours “well-capitalised buyers that have the market focus and operational expertise to extract incremental value” points to this dynamic.

Third, adjacent verticals, especially biologicals and animal genetics, have matured to the point where they can stand alone as investable categories. The biologicals segment is experiencing the fastest consolidation of any crop input category, driven by EU regulation, patent cliffs in synthetic chemistry, and shifting grower economics.

Agtech investment in 2025 was defined by what Verdant calls “pragmatic digitalisation,” with investors demanding proof of ROI and commercial traction, though it’s worth noting that Verdant’s Graeme McCracken is more cautious here. He observes that many AgTech categories still require additional scale, recurring revenue, and operating leverage before becoming attractive PE investment targets.

The buy side reveals where conviction capital sees durable value

Every unbundling creates a buy side. The buyers showing up for these assets tell us as much about the future of agrifood as the sellers do, and they fall into two distinct groups.

The first group is acquiring divested pieces from the conglomerate breakups. Balmoral Funds picked up Wilbur-Ellis’s nutrition division. GrainCraft acquired Bunge’s North American corn milling business. GDM bought AgReliant from KWS and Vilmorin. These are focused operators and financial sponsors stepping into the spaces that diversified companies have vacated.

The second group is taking out standalone businesses that fit a specific thesis, which is a related but distinct dynamic. DENSO Corporation, the Japanese automotive components giant, entered the vegetable seed industry with its acquisition of Axia Vegetable Seeds, a specialist in high-quality tomato seeds for greenhouses. This kind of entry by non-traditional players signals growing interest from outside the industry in acquiring differentiated breeding platforms and advanced R&D pipelines.

Abu Dhabi’s sovereign wealth fund ADQ entered exclusive discussions with Vilmorin & Cie for a 35% stake in its vegetable seeds business. Vilmorin delisted in 2023 and is now significantly connected to a sovereign fund with no previous link to agriculture.

In Australia, KKR acquired ProTen, one of the country’s largest poultry and agricultural infrastructure businesses, with over 700 broiler sheds. A financial buyer underwriting the long-term structural demand for animal protein through infrastructure assets backed by long-term contracts. In the same region, Cargill acquired full ownership of Teys Investments, consolidating its beef processing position across Asia-Pacific.

And Patient Square Capital completed a ~$4.1 billion take-private of Patterson Companies, a leading North American and UK supplier of dental and animal health products. The animal health distribution channel is attracting PE capital at scale.

Both dynamics point in the same direction. Whether they’re acquiring divested carve-outs or taking out independent companies, thesis-driven buyers are building focused platforms while legacy conglomerates simplify. The unbundling on the sell side is being met by targeted, conviction-driven buying on the other.

What I’m watching in 2026

Conglomerates are breaking apart, but within focused verticals, integration is still the winning logic

The most visible trend in Verdant’s data is ‘deconglomeration’: diversified players shedding unrelated divisions and retreating to core competencies. But the flip side of that story is equally important. Within those focused verticals, the appetite for capability-building acquisitions has not diminished. What’s changing is the unit of integration, from portfolio breadth to segment depth.

For founders, this distinction matters because the acquirers most active in 2025 were filling specific gaps in existing platforms. Understanding exactly where those gaps sit in your most likely acquirer’s stack is increasingly the work.

Biologicals M&A is approaching the moment that separates conviction from momentum

Biologicals attracted some of the most significant deal activity in 2025, and the strategic rationale from major acquirers sounded coherent. But now will be the real test of whether those acquisitions perform commercially at scale. The products need to prove consistent field results, the supply chains need to hold, and the bundled go-to-market strategies need to generate the revenue lifts that justified the acquisition multiples.

If that proof arrives, the category enters a self-reinforcing cycle where channel confidence rises, and farmer adoption broadens. If it doesn’t, expect a quieter deal environment as acquirers absorb write-downs and the next wave of biologicals startups finds fundraising harder.

The 2025 buyer universe is a spec sheet for what to build toward

Looking across the deals Verdant tracked, the acquirers with the most conviction and the clearest strategic rationale were paying for a fairly consistent set of attributes: proprietary data assets, defensible biological IP, demonstrated commercial traction in a specific crop or geography, and technology that slotted into existing workflows rather than requiring a full practice change.

For founders thinking about exit readiness over a five-to-seven year horizon, that list is more useful than any generic “build a moat” advice. The buyers who showed up in 2025 told what they were missing. The question is whether you’re building toward that gap.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Welcome to Issue #136 of Better Bioeconomy, weekly insights on how tech is reshaping food and agriculture for better human and planetary health. Thanks for being here!

BIO BUZZ

Products, partnerships, and regulations

🇺🇸 USDA and Department of War signed an MoU to elevate agriculture into the national security framework

• The agreement builds on the National Farm Security Action Plan announced in July, which targets agrifood research security, supply chain gaps, and domestic productivity. A new research partnership between DARPA and the USDA’s Office of the Chief Scientist has been formed, alongside a new Office of Research, Economic, and Science Security under the USDA.

• The MoU gives USDA agencies access to DARPA to accelerate the adoption of novel technologies. Its focus areas include biotech and biomanufacturing, American-made research, and building a “21st-century agro-defence workforce” through higher education programmes.

• The two departments will share intelligence on food system threats, technological requirements, and economic trends. USDA testbeds may be used to fast-track testing, validation, and deployment of agricultural biosecurity technologies through an “expedited but bounded pathway” that maintains safety and oversight.

Source: Green Queen

💰 Cellular agriculture infrastructure company IntegriCulture reaches profitability through a multi-sector strategy

• The Japanese company posted a net profit of JPY40M ($255k) for FY Oct 2024-Sept 2025. It funded itself through cosmetics and other non-food uses while continuing to supply cellular agriculture startups and run CRO (Contract Research Organisation) work.

• The company is transitioning from a supplier-CRO model to a CRDMO (Contract Research, Development & Manufacturing Organisation) while building full-stack infrastructure for cellular agriculture, including cell culture media, scaffolds, and low-cost hardware. Management says each segment is already profitable on a unit basis, with further growth expected in FY25 (Oct 2025-Sept 2026).

• Instead of relying on large, expensive stir-tank bioreactors, IntegriCulture focuses on compact packed-bed and adherent systems to reduce costs and risk. Its scale-up work is supported by Japan’s SBIR program, while scale-out development includes collaboration with Sumitomo Riko on low-cost, oxygen-infusion-free bioreactors.

Source: AgFunder

Thoughts 🤔:

IntegriCulture turning profitable is a useful reality check for cell ag. One of the first businesses to show durable economics is not a cultivated meat brand, but an enabler selling infrastructure, with cosmetics acting as the near-term profit pool rather than food.

Three signals stand out to me. First, “picks & shovels” are proving bankable before the headline products, echoing how CDMOs and toolmakers in biopharma captured value well ahead of many novel therapeutics. Second, diversification into non-food verticals is becoming the default de-risking pattern while food-grade regulation, CAPEX, and cost curves catch up. Third, their pivot away from massive stirred tanks toward modular, adherent and packed-bed systems hints that cell ag scale may look more like distributed solar than a few giga-factories.

This suggests that, at least in the near-term, the most resilient opportunities in cellular agriculture will sit in infrastructure, services, and adjacent ingredients rather than pure-play cultivated meat brands.

🌾 Syngenta becomes first to bring a commercially scalable hybrid wheat to market in Europe

• X-Terra hybrid wheat has received regulatory approval in France, with the first two varieties, SY Sphynx and SY Xanthis, available to farmers for the 2026 sowing season. The milestone is the result of more than 15 years of work cracking wheat’s genome, which is nearly six times larger than the human genome.

• The varieties are engineered for climate resilience, offering improved tolerance to septoria (one of Europe’s most economically damaging wheat diseases) and integration with Syngenta’s Cropwise digital platform for early disease-pressure alerts. The approach positions X-Terra as a system-level innovation rather than a standalone seed product.

• Syngenta plans to extend the platform to the UK and Germany following the French approval, framing hybrid wheat as a cornerstone of Europe’s next generation of climate-resilient cereals. For a crop that supplies nearly one-fifth of global calories, the shift to hybrid performance carries major food security implications.

Source: AgTechNavigator

Thoughts 🤔:

Wheat might be is starting to follow maize and barley toward a world where hybrids become the performance benchmark. That can change who captures value, because more of the category shifts from farm‑saved seed to annual repurchase tied to proprietary genetics and playbooks.

Bundling X Terra with Cropwise and integrated agronomy reframes “seed choice” into “system choice.” Once your yield risk, disease alerts, and input protocols are tuned around a specific hybrid and data stack, switching becomes costly.

The question is whether the economics clear the bar for European growers after the seed premium and tighter management recipes. If the hybrid‑plus‑digital package can reliably buy a few points of margin and more predictable yields under climate volatility, this looks like the start of a cereals platform shift.

🔎 Carbon Robotics unveils its Large Plant Model, the ‘most advanced AI model’ for plant detection and identification

• The Large Plant Model (LPM) is trained on years of plant images across varying soil types, climates, and growth stages. It can identify a new weed immediately, including how it will look from different angles and as it matures, without sending data back to a data centre for retraining. Farmers with Carbon Robotics’ existing LaserWeeder machines can access it via a software update.

• The new PlantProfiles feature lets farmers use the Carbon Robotics iPad Operator App to select photos of their specific crops and weeds. The system then automatically adjusts to the parameters of that field, cutting setup time and making the tech usable without deep technical knowledge.

• LPM moves away from the traditional computer vision architecture used in earlier LaserWeeder models, which could not adapt to new plants in real time. CEO Paul Mikesell says parts of the AI architecture are now more advanced than those used in self-driving cars, and positions LPM as bringing that frontier AI capability to a part of agriculture that has largely missed out on recent advances.

Source: AgFunder

Thoughts 🤔:

This is showing up as three constraints tighten at once: herbicide resistance keeps rising, input economics remain volatile, and seasonal labor windows are still unforgiving.

Cloud-centric ag AI also hits practical ceilings in the field, since connectivity is uneven and in-row decisions cannot wait on round trips to a data center. Carbon’s bet is edge inference plus a fleet data flywheel, turning deployed machine-hours (and a very large labeled plant dataset) into compounding model advantage.

The playbook rhymes with AV autonomy and precision equipment: ship useful hardware, harvest messy edge cases at scale, then shift value capture toward the perception and decision layer. The question is whether the software stack becomes the thing growers renew and competitors struggle to replicate.

If Carbon can keep performance strong across crops and regions, this shifts the advantage toward whoever has the most machines working in real fields.

🍄🐄 Joyn Foods earned K-12 approval in South Carolina, allowing it to sell its mycelium-enhanced blended meat to schools across the state

• The US startup (formerly Mush Foods) secured statewide purchasing access through the South Carolina Purchasing Alliance (SCPA). This approval is important because school bids require strict compliance with USDA standards, competitive pricing validation, and proven student acceptance through pilot tastings.

• Its 50Cut blend mixes beef, mushrooms, and mycelium. By partially replacing conventional beef, the blended protein reduces environmental impact while improving nutrition without asking school kitchens to change how they operate.

• In tastings, students gave the meatballs a 96% approval rating and the burgers 92%. Even K-2 students responded well, with 84% positive feedback. With the National School Lunch Program serving nearly 30 million meals a day, even small formulation changes in ground beef can impact millions of meals annually.

Source: Green Queen

Thoughts 🤔:

One thing I keep coming back to in “alt protein adjacent” stories is that the real bottleneck is rarely the ingredient science. It’s distribution through procurement systems and kitchen reality.

In the near term, scale tends to flow to products that fit existing institutional rails. Bid lists, meal pattern compliance, and predictable pricing matter more than consumer branding when the buyer is a school district, hospital, or university. If something can drop into the same recipes, equipment, and prep flow, it meaningfully lowers the adoption cost.

Blended formats also feel like a pragmatic cultural layer. They can improve nutrition and reduce footprint without asking customers to opt into an identity statement. In a polarized food narrative, that “balanced protein” positioning might end up being a feature. This is especially true in public and contract catering, where quiet reformulation can move faster than consumer sentiment.

🍼 Tech breakthrough lets PFerrinX tune iron loading in its precision-fermented lactoferrin by application

• The startup has developed a process to precisely control iron saturation in its NanoFerrin ingredient, producing three distinct profiles: apo-lactoferrin (0% iron) for immunity and cosmetics, native lactoferrin (~10%) for infant nutrition, and holo-lactoferrin (100%) as a gentle iron supplement.

• Most precision-fermented lactoferrins, the company claims, default to high iron saturation (75%+), which limits their utility in applications like infant formula, where lower levels are preferred.

• PFerrinX has secured manufacturing capacity of 200 metric tons across sites in the EU and Asia, with room to scale to 400 tons based on demand. The company says it can currently produce at cost parity with dairy and plans to undercut dairy’s $300/kg production cost within 24 months, with multiple regulatory filings underway.

Source: AgFunder

Thoughts 🤔:

The first wave of animal-free lactoferrin was a pure supply story: can we make the same molecule, at scale, without cows. The PFerrinX news points to a next phase: spec-defined ingredients where attributes like iron saturation are tuned, validated, and sold as the product itself rather than left as a back-end QC detail.

That could shift the competitive moat in precision fermentation downstream. Strain IP and stainless steel still matter, but differentiation starts to look like finishing, QA, and the ability to guarantee a functional state batch after batch, and eventually across sites. It’s a more “formulator-first” world where customers brief against performance tolerances, claims defensibility, and risk profiles.

Once buyers have options, procurement behavior tends to reward suppliers who can industrialize reliability: tighter variance, better comparability packages, clearer spec sheets, fewer surprises during scale-up.

🍫 Cargill has partnered with Kokomodo to validate commercial cell-based chocolate solutions, with support from EIT Food

• The partnership builds on EIT Food’s €20,000 investment in Kokomodo in late 2024 and its selection as one of 16 startups in the 2025 RisingFoodStars cohort. These programmes are designed to help promising agrifood startups validate, scale, and commercialise transformative tech more rapidly.

• Cargill and Kokomodo will run trials in industrial settings, testing how cell-based cocoa performs in beverages, dairy, and confectionery. They’ll assess functionality, taste, and whether it can be produced at scale. Initial functional test results are expected in the coming months.

• Kokomodo’s tech uses cocoa cells sourced from premium Central and South American beans to create sustainable, future-proof cacao ingredients. Cargill says the project fits its “portfolio mindset,” giving product developers another route for cocoa formulations while reducing pressure on traditional supply.

Source: Green Queen

BIO BUCKS

Funding, M&As, and grants

🧬 Resurrect Bio raised $8.1M Series A led by Corteva to advance its AI-powered platform for restoring crops’ innate disease resistance

• UK-based startup has developed a three-part platform, anchored by its FloraFold AI tool, that maps the molecular interactions between pathogen effector proteins and plant NLR immune proteins. By identifying precisely where pathogens bind to suppress a plant’s immune response, it can pinpoint the exact gene edits needed to block that interaction and restore full resistance to fungal disease, oomycetes, and nematodes.

• Rather than performing gene edits itself, Resurrect Bio delivers validated editing targets to seed company partners who apply techniques like CRISPR to their own germplasms. The approach is crop-agnostic and pathogen-broad, with current focus on soy, corn, wheat, rice, cotton, and brassicas. Some resistance changes require as little as a single amino acid substitution.

• The company is targeting 18 disease resistance traits across its blockbuster crops in the next three years, alongside improvements to FloraFold’s predictive accuracy. Several joint development agreements with leading seed companies are expected to be announced shortly. Revenue will come from milestone payments, success fees, and profit-sharing on licensed seed traits.

Investors: Corteva Catalyst, Calculus Capital, Pymwymic, UKI2S, SynBioVen, and AgFunder

Source: AgFunder

Thoughts 🤔:

The few signals I’m watching in the Resurrect Bio news are about where trait discovery is trying to go next. First, the industry seems to be testing a shift from bespoke, single-trait programs toward horizontal platforms that can repeatedly identify edit targets across crops and pathogens. The moat moves from “we have one great trait” to “we have the fastest, most reliable discovery-to-deployment loop with seed partners.”

Second, the emphasis on minimal edits, looks like a deliberate regulatory and adoption strategy. It’s a way to fit through the narrowest NGT‑style gates and keep optionality in tough markets like the EU and UK, while also making the edits easier to explain to downstream buyers.

Third, Corteva leading via Catalyst reads like pre‑bought option value ahead of its planned seed and crop protection separation in the second half of 2026. In a split, each entity needs its own pipeline narrative. External trait platforms can support the seed side’s cadence and give the crop protection side a path into more durable, lower‑scrutiny resistance “products” as chemical pressure builds.

🤖 Upside Robotics raised a $7.5M seed round to cut fertiliser use and waste in corn crops

• The Canadian startup builds lightweight, solar-powered robots that apply fertiliser only when plants need it. The company says the approach can cut fertiliser use by 70% and save farmers about $150 per acre per season. It uses its proprietary algorithms with soil and weather data to optimise nutrient delivery

• The startup has shown rapid traction, expanding from 70 acres in 2024 to 1,200 acres in 2025, with projections to exceed 3,000 acres in 2026 while keeping 100% customer retention.

• Only 30% of traditionally applied fertiliser is absorbed by crops, leaving 70% wasted. Upside targets this inefficiency, starting with corn, one of the most fertiliser-intensive crops. With more than 200 farms on its waitlist, Upside plans to use the funding for R&D, to scale operations, and to expand into the US corn belt.

Investors: Plural, Garage Capital and the founders of Clearpath Robotics

Source: TechCrunch

🎯 Kilter raised €6.5M pre-Series B led by Kubota to scale its autonomous spot-spraying robot for high-value crops

• The Norwegian startup’s AX-1 spraying robot identifies individual weeds using AI and applies only a few droplets of crop protection product to a 6×6 mm target area. The system avoids damage to surrounding crops while significantly cutting input volumes. It currently supports more than 15 crop types.

• Alongside the investment, a distribution partnership with Kubota will make the AX-1 available through dealer networks in Germany and the Netherlands from 2026. The arrangement pairs Kilter’s precision nozzle tech and AI capabilities with established machinery expertise and commercial infrastructure across key European markets.

• The robot is positioned for growers facing herbicide resistance and tightening regulatory frameworks in Europe, where effective active ingredients are becoming increasingly scarce. The new funding will support international expansion, further tech development, and pilots across additional crops and markets.

Investors: Kubota, SBG Invest AS, Pymwymic, and Nufarm

Source: Future Farming

Thoughts 🤔:

Kubota leading Kilter’s round looks like an OEM testing a new role: platform orchestrator in crop care. The interesting delta is the pairing of capital with a Smart‑Farming‑branded, dealer‑led rollout in markets like Germany and the Netherlands. That effectively turns distribution, servicing, financing, and uptime into the core product.

If that model works in specialty crops, it could become the fastest path for robotics to clear adoption friction, because growers buy reliability and support as much as they buy capability.

The other interesting shift is how “precision” is being reframed. Ultra‑targeted spot spraying is increasingly regulatory and operational de‑risking, not just an input‑saving feature. In Europe, where crop protection constraints tighten, how product was applied could become a compliance edge and a buyer requirement.

🐝 Polybee raised $4.3M seed round to scale its drone-based pollination and AI yield forecasting tech

• The Singapore-based startup uses small drones that launch and recharge on their own. Fitted with cameras, they track crop health, ripeness, and variability throughout the season. That stream of images feeds models that tell growers where and when to harvest to improve yield and margin.

• Customers report up to 3x higher profits from better harvest timing and earlier detection of crop stress. In greenhouses, autonomous pollination has delivered up to 15% yield gains. These benefits are especially valuable during shoulder seasons when market prices are highest.

• The new funding will help Polybee scale quickly, aiming for a five-fold jump to more than 4,000 acres by 2026. The company sells an end-to-end subscription (fixed fee per hectare, no upfront drone purchase) that cuts labour, improves consistency, and reduces the risk of pollination versus bumblebees or manual work.

Investors: Paspalis Capital, elev8 VC, SEEDS Capital, and strategic angel investors, including Blue River Technology founder Jorge Heraud

Source: AgFunder

Thoughts 🤔:

The interesting signal in Polybee’s approach is the reframing of pollination as an operational risk line item that can be engineered and monitored. Bees and manual stimulation work, but they introduce labor exposure and biosecurity complexity. If controlled airflow can deliver consistent fruit set with lower variance, pollination starts to look like fertigation or climate control: a managed input with tuning knobs, logs, and accountability.

Another interesting signal is the move from maximizing yield to maximizing timing. “Right-time yield” is basically revenue management for produce. If you can predict ripeness and variability continuously, you can make better harvest decisions, hit contract specs more reliably, and capture shoulder-season pricing without guessing.

This could be a real step toward algorithmic biology in high-value crops if unit economics and reliability hold across sites. Otherwise it stays a premium tool for labor-constrained, biosecurity-sensitive growers.

❄️ CryoBio raised $1.3M in pre-seed funding to cut frost-related crop losses with a biologic crop protectant

• Frost damage drives an estimated $30B in global food waste, enough food to feed 200 million people. The New York-based startup was created to fill a market gap in frost-specific crop protection solutions. Frost damage has doubled over the past decade and is expected to double again as climate change and earlier planting push growers into riskier windows.

• The company’s lead product is an anti-freeze molecule made through precision fermentation. It can protect crops from frost for several days up to two weeks, depending on the crop. CryoBio is also working on a seed-treatment version designed to run through farmers’ existing equipment.

• Aiming for a 2028 launch, CryoBio is running five-acre field trials with nearly 30 growers across New York, Maryland, and Washington. The next round will fund biomanufacturing scale-up from grams to kilograms and expand trials to larger acreage.

Investors: Marble, Ag Ventures Alliance, NY Ventures, Launch NY, FuzeHub, and impact angels.

Source: AgTechNavigator

Thoughts 🤔:

Frost is becoming first-class climate risk category with its own solution stack. We now have credible estimates that spring frosts destroy food equivalent to feeding hundreds of millions of people each year, and a new wave of biological tools is targeting that loss. That reframing is important because once a risk is quantified and repeatedly priced into losses, it competes for the same mental and budget real estate as drought, heat, and disease.

The big question is whether performance holds across messy field conditions, not just in the best case. Biological and bio based claims often compress when you add region, variety, and microclimate variance. If multi-site data shows consistent ROI, frost could become a budget line item rather than an emergency response.

🦠 MicroHarvest secured a €5.5M German government grant to build an industrial-scale microbial protein plant in Leuna

• The plant will produce 15,000 tonnes of protein annually for pet food, human food, and aquaculture. Production is expected to begin within two years, with molasses as its primary feedstock through an already-secured regional supply chain.

• The Hamburg-based startup’s biomass fermentation process runs in just 24 hours and produces protein with over 60% raw content, a complete amino acid profile, high digestibility, and an umami flavour that improves palatability. Its MPX Care pet food ingredient has already appeared in The Pack’s Gut Bites and Vegdog’s dog and cat treats, while MPX Boost and Hilix target the aquaculture market.

• The company reviewed 40 European sites before selecting Leuna for its feedstock proximity, industrial infrastructure, and surrounding agri-processing network. MicroHarvest will work with Industriepark Leuna as its park operator and utilities provider, with local and federal stakeholders supporting the project through permitting and investment incentives.

Source: Green Queen

🐄 Ruminant BioTech launched CAD$7.6M program to validate its slow-release methane-reduction bolus for pasture-raised cattle

• The New Zealand startup’s slow-release bolus sits in the animal’s stomach and steadily delivers bromoform, a methane-inhibiting compound. Unlike feed additives suited to intensive farming, this approach works in grazing systems where daily controlled feeding is not practical. Animal trials suggest that one dose delivers a 75% reduction in methane for over 100 days, with the team targeting 6 months.

• The three-year, on-farm program is designed to secure Canadian registration for beef cattle and generate data needed for future expansion into the US, Brazil, and the EU. It is backed by a CAD$2.8M Alberta government grant, CAD$3.4M from two industry partners, and CAD$1.4M from Ruminant BioTech’s own funds.

• The company’s business model leans on carbon insetting and offsetting to cover costs where there is no direct productivity payback for farmers, pointing to carbon markets now worth over $1T and covering 25% of global emissions. CEO cited Brazilian meat giant Marfrig’s methane targets as evidence that major food companies are seeking solutions.

Source: AgFunder

BIO READS

Thought-provoking opinions

🇮🇳 India will see IPOs in agritech before America or Europe

• Omnivore’s Mark Kahn argues that Indian agritech is structurally different from Western markets and could see IPO exits earlier, given more accessible public markets and strong domestic investor appetite.

• Mark traces India’s agritech boom to the post-2016 shift from consumer e-commerce to B2B marketplaces and farm-to-consumer brands. Fragmented smallholder supply chains have been reorganised through service-based and asset-light models.

• He also points to growing opportunities in biomanufacturing, full-stack CDMOs, robotics, and climate-aligned companies. And he rejects the idea that smallholder agriculture is charity-driven rather than a scalable, tech-enabled business opportunity.

Source: Mark Kahn

🧩 Antifragile agriculture: How to stop playing the old game

• Christine Gould, founder of GIGA Futures, argues that agrifood innovation is structurally fragile because incumbents, investors, and policymakers continue to optimise for protecting the existing model rather than building new growth engines.

• Drawing on examples from sectors such as defence, tech, and consumer goods, she introduces the concept of “antifragile agriculture.” In this approach, volatility is treated as a catalyst for capability-building through portfolio thinking, scaling infrastructure, and circular innovation systems.

• The piece is both a critique of risk-averse capital and corporate behaviour, including signals like the shutdown of FMC Ventures, and a call for deliberate architectural shifts that would allow agriculture to generate its own AWS- or ZX-style breakout platforms.

Source: Christine Gould

🤔 The ag input industry is entering its most active restructuring cycle since the 2010s consolidation wave

• The global crop input industry is undergoing a structural reset driven by patent expiries, generic competition, regulatory pressure, and higher capital costs.

• After a decade of consolidation in the 2010s, large incumbents such as FMC, Syngenta, BASF, Corteva, Bayer, UPL, and Nufarm are now reassessing their portfolio composition, capital structures, and organisational designs.

• Jorge argues that what we are seeing is not isolated corporate events, but a broad cycle of unbundling and strategic repositioning that could reshape competitive dynamics across seeds, crop protection, and biologicals over the next several years.

Source: Jorge Fernández Vidal

🏆 The FoodTech World Cup is officially underway for 2026!

Applications are now open for Founders pushing the boundaries of biotech in food.

FoodHack and Nestlé Research are ready to uncover fresh talent and spotlight the most promising solutions.

How it works: 15 top teams will advance through to the virtual Semi-Final before the 5 brightest go head-to-head at the HackSummit’s Grand Final in Lausanne on 23rd April.

Already confirmed on the Jury are Bodil Sidén (Kost Capital), Celine Schiff-Deb (Mista / Givaudan), Gil Horsky (Flora Ventures), Eugenia Barcos (Nestlé Research), Steve Molino (Synthesis Capital).

Enter your startup today!

If you found value in this newsletter, consider sharing it with a friend who might benefit! Or, if someone forwarded this to you, consider subscribing.

Share

Subscribe now

Thank you, and have a great day!

Disclaimer: The views and opinions expressed in this newsletter are my own and do not necessarily reflect those of my employer, affiliates, or any organisations I am associated with.

Thanks for being here. For Issue #135 of Better Bioeconomy, I sat down with Ferry Kamp, Founder and Managing Partner of Remagine Food.

Ferry brings over two decades of food industry experience, with senior leadership roles at Unilever, leading corporate venture building for clients such as Nestlé, Danone, and Heineken at Startupbootcamp-Innoleaps. He has driven commercial scaling and international expansion at plant-based pioneers, including Meatless Farm and GREENFORCE, and now backs and builds companies through Remagine Food, an operator-led food system investment platform that replaces passive capital with hands-on scale execution.

In our conversation, we unpacked what “invest and build” means in practice, why so many food tech startups fail to scale, how to design corporate partnerships that help startups, and how to match capital types to the risks in food tech. We also dug into portfolio construction when exits are mostly strategic M&A, and why Ferry is leaning into system-level infrastructure rather than premium sustainability plays.

Let’s jump in!

Inside an operator-led food system fund

Remagine Food began as a hands-on venture builder that helped food tech companies with commercial scaling, restructuring, and growth for other investors’ portfolio companies. Over time, Ferry and his partners kept running into the same challenge: before they could actually accelerate growth and scale, they first had to spend months fundraising just to secure the budget to do the work. So as a natural next step, they joined forces with Hive Capital to combine building and investing, a model designed to accelerate scale, not just fund innovation.

In practice, operator-led for Remagine Food means they “don’t stop at the cheque.” When they invest, they also commit time, people, and execution capacity. That includes hands-on work on commercial strategy and pricing, support on partnering with manufacturers and ingredient suppliers, help structuring follow-on rounds and M&A pathways, and, in some cases, stepping in as interim operators or board-level executors.

A traditional VC might invest in over 20 companies and expect one or two to break out, but in food, capital cycles are long, scale-up is messy, and most failures come from commercial and operational friction. Ferry’s view is that it is better to back fewer companies and be deeply involved in how they grow, partner, and finance the journey from idea to scale.

The classic VC playbook is rooted in software and Silicon Valley, where low marginal costs and fast feedback loops make it rational to swing often and let compounding take care of the rest. Food is different. The industry is more complex, with many actors from farmers to processors to retailers, and the path to scale runs through manufacturing, regulation, and supply chains that are unforgiving. In that environment, Ferry argues, you need investors who know the industry and are willing to behave like operators, not just capital allocators.

Why 90% of food tech fails to scale

The world does not lack food tech innovation. Technologies exist, and solutions are ready, but they lack scale and real impact. Many products are founded with the ambition to improve health, the environment, or social outcomes, but without scale, they never make a meaningful impact.

Ferry sees five recurring failure modes:

1. Commercial naïveté: Teams spend five or six years in the lab, then assume they can go straight from the lab to the shelf and are surprised when customers are not lining up. They have brilliant tech but no real buyer insight, no clear go-to-market, and no robust pricing model.

2. Scale illusion: Pilot success is mistaken for scalable demand. A product that works in one pilot or a small market is assumed to be ready for global rollout, without the intermediate steps of small-scale launch, national or regional rollout, and proof of a repeatable sales model.

3. Manufacturing mismatch: Technology that works in the lab does not work on industrial lines, or cannot be brought to a cost level the market will tolerate. The result is a solution that is technically sound but trapped in a small premium niche, far from the total addressable market (TAM) promised in the early pitch deck.

4. Capital structure errors: Teams use high-risk VC capital to fund heavy CAPEX, long manufacturing ramps, or low-margin infrastructure that should be financed with debt, strategics, or customers. The wrong capital ends up carrying the wrong risk.

5. Founder role lock-in: Technical founders with strong science backgrounds often stay too long in commercial leadership roles they are not equipped for, becoming bottlenecks as the company shifts from invention to scale. Ferry sees a clear difference between first-time founders, who often have blind spots around their own gaps, and serial founders, who know when to bring in complementary leadership or step aside.

Given that context, Remagine Food’s evaluation lens goes beyond “Is the tech good?” and asks three harder questions:

1. Is there a repeatable buyer with urgency? Ferry wants to know who needs this at scale now and why, and whether they are willing to pay today.

2. Does the company understand its “scale unit”? That could be tonnes per year, factories per region, or customers per line, and can articulate how that unit grows without breaking unit economics.

3. Is the founding team adaptable? The best founders evolve from inventors to builders to leaders. They pivot if needed and adjust the business model as the market reacts.

The top 10% of companies that break out are not always the most innovative, but they are the most adaptable.

Building corporate partnerships that work for startups

On paper, startups and corporates look like a perfect match. The speed and creativity of a startup, paired with the scale and muscle of a multinational. In reality, Ferry has watched many of these collaborations unravel into culture clashes and stalled pilots. Startups optimise for learning and iteration, while corporates show up asking about gross margins and risk controls from day one.

His view is that most collaborations are designed for the corporate, not the startup. Corporate processes and governance quickly become straitjackets that slow founders down, drain teams, and sometimes push them out. If a partnership does not remove the startup’s biggest scaling bottleneck, it is a distraction.

From a founder’s seat, good partnerships share four non-negotiables:

1. Named internal owner: Someone inside the corporate whose P&L or bonus depends on the partnership’s success. If everyone “supports” the startup but no one owns it, nothing moves.

2. Clear value exchange: Not vague “learning” or “exploration,” but explicit access to assets: manufacturing capacity, customers, regulatory muscle, or distribution.

3. Time-bound decision path: Founders should know what decision gets made in six to twelve months if milestones are hit, whether that is a rollout, a strategic investment, or more capacity.

4. Founder protection: IP ownership, decision rights, and commitments need to be clear up front. The agreements Ferry structures with strategics include in-kind contributions like production, clinical trials, and customer access to avoid the “no one picks up the phone” problem after the deal closes.

Finding the right capital stack for scaling

Food tech gets into trouble when teams try to force it into a software-style funding approach. Ferry’s golden rule: Match the capital type to the risk you’re trying to remove. Most startups fail because they fund scale risk with innovation capital.

He breaks the stack into four broad buckets.

Venture capital

VC should fund early tech validation, building the core team, and first commercial proof points. This is the phase where high-risk capital is needed because the basic question “does it work and will anyone pay for it” is not yet answered, and banks or strategics will not step in. Where Ferry sees trouble is when VC money is used for heavy CAPEX, long manufacturing ramps, and slow-margin infrastructure.

Strategic capital

Money from large food or ingredient companies is best used once there is clear commercial pull and should buy more than cash: manufacturing access, raw materials, go-to-market acceleration, and de-risked scale.

In one precision fermentation deal Remagine Food is closing, two of the largest ingredient companies are investing capital and in-kind support, including production, clinical trials, and customer introductions, with automatic follow-on investment if the startup hits its milestones. Those strategics are also likely M&A candidates down the road, effectively building a highway to scale rather than a series of disconnected rounds.

Debt and project finance

Debt becomes viable when unit economics are proven, contracts or offtakes exist, and CAPEX has predictable returns. At that point, the business starts to look like something a bank or project financier recognises: evidence of demand, line of sight to payback, and enough operational stability to carry leverage.

Grants

These are powerful tools for technical risk and early pilots, especially in ecosystems like the Netherlands, where clusters around precision fermentation or alternative proteins offer structured support. The risk is that they delay exposure to real customers, pushing teams to optimise for winning proposals instead of winning buyers.

Designing portfolios for strategic outcomes, not unicorn IPOs

In a market where liquidity is scarce and most exits are likely to be strategic M&A rather than IPOs, portfolio construction warrants reconsideration. For Remagine Food, in collaboration with Hive Capital, that starts with treating the fund itself as a collective: bringing strategics in as LPs, working with governments and other investors, and using that ecosystem to see what the industry wants solved.

That information then drives capital allocation. If you know what large buyers are trying to solve, you can build or scale companies against those specifications, increasing the odds they become customers, strategic investors, or acquirers. Strategics are also more patient than purely financially driven investors, giving companies time and technical and commercial support, which Ferry believes leads to stronger businesses and higher valuations.

For fund design, this means underwriting to credible buyers, not just big TAM numbers, and accepting that “base hits” matter. Multiple 2-5x exits into strategic buyers can drive strong fund returns if you own meaningful stakes and shorten growth cycles through active involvement. Food tech does not need unicorns to work. It needs repeatable strategic outcomes mapped against clear M&A pathways.

What’s interesting now and what isn’t

Asked what excites him, Ferry leans toward system innovation rather than single products. Four clusters stand out in his mind:

1. Precision fermentation platforms with credible pathways to reduce costs.

2. Ingredient technologies that plug into existing supply chains.

3. Fermentation, enzymes, and process technologies that upgrade today’s food system rather than replace it.

4. Food waste valorisation: turning food loss into high-value ingredients and building profitable businesses.

Food loss is the topic the conversation kept circling back to. More than 40% of food produced is lost or wasted, and roughly half of that loss happens early in the chain at farms, in transit, or in factories. That early-stage loss is worth hundreds of billions of dollars and represents one of the biggest margin levers available to producers and retailers.

Yet loss and waste are still treated as background noise, simply baked into business cases. Ferry expects prevention to move from nice-to-have to must-have as it becomes mathematically impossible to feed a growing population while throwing away 40% of output, and as companies see that cutting losses improves margins more than marginal logistics tweaks.

On the flip side, there are areas where Remagine Food is deliberately cautious (for now). Ferry is less interested in end-of-chain brands and propositions that rely on consumers paying a premium for sustainability. After working closely with alternative protein and meat analogue companies, Ferry no longer believes that direct meat imitation is the main path forward, especially when plant-based burgers are priced much higher than meat for households that rarely cook burgers in the first place.

He sees more potential in natural, clean and flexible plant-based ingredients that slot into dishes people already cook, at budgets they can live with. That lens also shapes his view of cultivated meat and some alternative protein plays: they are not dead, but they are ahead of their financing reality, with unresolved questions around consumer adoption and reaching mainstream price points before other solutions fill the space.

Ferry’s closing thoughts: “Food system transformation will not be led by better pitch decks. It will be led by people who know how to build, finance, and operate in complexity. That is the gap Remagine Food is trying to close.”

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

This issue marks three years since I started the newsletter! 🎂

I didn’t expect it to become such a meaningful part of my life, but it has. Writing each week (well, almost. I’ve managed ~87% consistency, because some weeks life gets in the way) has sharpened how I think, introduced me to cool people I wouldn’t have met otherwise, and opened doors I am very grateful for. It has, without a doubt, changed the trajectory of my life.

Thanks for reading, replying, sharing, and tolerating the bad takes and cringey memes. I’m excited to keep going.

Okay, now back to work. Let’s dig in! 🍽️

BIO BUZZ

Products, partnerships, and regulations

☕ Gates Foundation-backed pilot equips East African coffee farmers with Helios AI to manage climate risk

• Helios AI and the Committee on Sustainability Assessment (COSA) are running a Gates Foundation-funded pilot in East Africa to help coffee farmers make better calls on production and sales. The project delivers weather and market forecasts meant to reduce risk from climate variability and price volatility.

• A core goal is data fairness: farmers should benefit from the data they generate. By testing which forecasts and insights matter most, the initiative aims to make climate information practical and usable at the cooperative and farm level.

• Helios is adapting technology traditionally used by global commodity buyers to support smallholder farmers. The platform covers more than 75 commodities across 90 countries, and this pilot shows how the same AI tools can serve both agribusiness giants and small farms.

Source: AgTechNavigator

Thoughts 🤔:

Coffee outcomes often come down to who gets good information first. Big buyers usually spot weather and price risk early, then adjust buying. This pilot is interesting because it tries to put that “early warning” in the hands of cooperatives and see if it changes day-to-day decisions like when to pick, when to sell, and how to plan together.

The catch is that information alone rarely shifts the price farmers get. If cash needs, lack of storage, or limited buyer choice force quick selling, the forecast doesn’t become leverage.

So the question is: does the signal connect to action? That will depend on clear playbooks and practical tools like shared storage, short-term credit, and more routes to market.

👀 MrBeast ate cultivated chicken from Upside Foods on camera

• It marked a major visibility moment for cultivated meat. With 460 million YouTube subscribers, his on-camera experience put the category in front of primarily Gen Z and Gen Alpha audiences (groups that already engage with food as content and identity).

• The lab visit also helps weaken the “lab-grown frankenfood” stigma. Watching him ask questions and taste the product frames cultivated chicken as 'normal, approachable, and worth trying.

• For Upside Foods, this is cultural validation with real commercial implications. Even a small wave of consumer interest can shift conversations with restaurants, retailers, and major protein players who still doubt customers will ask for it.

Source: Green Queen

🧠 Cropin launched an AI-driven agrifood platform aimed at reducing risk across the global food value chain

• Called Cropin Ecosystem, the plug-and-play offering draws on more than a decade of agricultural data and advanced AI models to support decision-making from sourcing through supply planning. Cropin claims its models deliver over 90% forecasting accuracy, supporting supply assurance and planning for pests, diseases, and price swings.

• The platform is built to help companies manage today’s biggest agricultural pressures like climate volatility, geopolitical uncertainty, and supply chain disruption. Better visibility into farms and supply can steady availability, protect margins, support climate-smart farming, and help meet tighter traceability and sustainability rules.

• The company says customers can deploy the system quickly and see operational changes within six months. Cropin also points to partnerships with cloud providers, AI and IoT firms, universities, and NGOs as part of its longer-term approach to food-system stability.

Source: AgNewsWire

Thoughts 🤔:

Cropin looks to be repackaging existing ag intelligence into a partner-bundled ecosystem and selling implementation certainty, with claims of sub-six-month deployment and high model accuracy. That six-month promise is a go-to-market tell for buyers under margin pressure who want short payback windows and lower integration risk.

Prediction outputs are being wired into procurement decisions. If they are reliable enough to change contracting terms, hedging posture, and inventory buffers, advantage shifts to whoever owns the data plumbing and workflow integration. If not, this risks becoming a services-heavy integration story where “ecosystem” is mostly packaging.

🍬 2nd Nature applies AI to discover clean-label sweeteners hidden in agricultural side streams

• The Cincinnati-based startup will start commercial sampling this quarter. First up are natural, non-caloric sweeteners and umami flavour enhancers meant to help food manufacturers cut sodium. The ingredients are sourced from abundant agricultural side streams.

• Its AgWaste Portal runs several AI models like taste prediction, structure-based screening, and sequence-similarity searches to identify candidates faster and improve hit rates. The company pairs discovery with scalable extraction know-how.

• Early sweetener candidates are 50-100× sweeter than sugar, clean-tasting, and non-caloric. Because they come from low-cost side streams, the company says they could compete on cost with stevia or monk fruit.

Source: AgFunder

Thoughts 🤔:

2nd Nature shows that ingredient innovation is not always about finding the next rare plant. It’s about pulling value out of what the food system already moves in bulk. When your inputs are commodity byproduct streams, the edge is taking messy, variable feedstocks and turning them into tight specs at a cost that holds.

If this model works, leverage comes from two core assets: steady access to high-volume side streams, and process know-how that standardizes them. Their partnerships with processors, CDMOs, and ingredient suppliers point to the real constraint. Manufacturing and distribution ties can often set time to market more than discovery.

🥛 One.bio allows formulators to add “invisible” fibre to foods and beverages without affecting taste or texture

• The California-based startup does this by breaking polysaccharides into soluble, palatable oligosaccharides while keeping the fibres’ functional structure. One.bio is reframing fibre as a family of ingredients (not a single nutrient), and connects fibre structures to microbiome and metabolic outcomes using a database of 4,000+ characterised fibres and a high-throughput fermentation system.

• The company plans to launch first through GoodVice, its in-house consumer brand. Early prototypes include a sparkling seltzer with 20g fibre and a chocolate milkshake with 20g fibre and 20g protein.

• In a 63-person, two-week clinical study, participants consumed up to 20g/day of one.bio’s oat fibre reported better GI comfort, improved glucose control, and better mood and focus.

Source: AgFunder

Thoughts 🤔:

I like one.bio’s framing of fiber as a programmable input. Start with the outcome you want (eg: more butyrate) then pick the fiber structure and dose that produces it consistently across people. Platforms that own the structure-to-response map and the evidence package brands can use, will have the edge.

The timing is right in the GLP-1 moment. Consumers and brands are increasingly focused on satiety, glucose stability, and gut tolerance, and fiber is a plausible food-based tool for addressing those outcomes. The bar should be clinical data, and it seems like one.bio is moving in the right direction.

🤰🏽 How Helaina is redesigning women’s nutrition from one size fits all to life stage-specific biology

• Women’s nutrition has historically been misdesigned, with most legacy ingredients selected for convenience rather than biological relevance. As a result, many women experience tolerance issues, inconsistent outcomes, and formulas that don’t fit the shifts of menstruation, pregnancy, and menopause.

• In 2025, several women’s wellness brands launched products using effera, Helaina’s precision-fermented human lactoferrin. Effera is bioidentical to human lactoferrin and is designed to help regulate iron, support gut and immune function, and support metabolism and the microbiome. Because it matches human protein structure, it is better tolerated and more consistent than bovine or other animal-derived options.

• Helaina’s bigger move is to design products and run studies by life stage. That means formulations and clinical studies built around real physiological seasons like menstruation, pregnancy and postpartum, perimenopause, menopause, athletic training, and periods of high stress.

Source: Green Queen

Thoughts 🤔:

Credit to Helaina for pushing an evidence-first approach. recision fermentation keeps improving at producing human-sequence proteins, and as more startups learn to make the same molecule, robust evidence starts to become the differentiator.

Helaina is leaning into women-specific studies with pre-specified endpoints that brands can reference in compliant claims. If the results are repeatable in clearly defined cohorts, defensibility shifts away from strain or process IP and toward endpoints, tight protocols, and a mechanism that holds up under scrutiny.

That’s where standard biomarker panels start to matter as legit category infrastructure. They make outcomes comparable, reduce ambiguity, and de-risk procurement decisions for brands.

🐄 Aleph Farms partnered with CDMO Cell Agritech and set up an entity in Singapore to serve as its APAC hub

• The move fits the Israeli company’s asset-light approach to scaling cultivated meat, relying on local contract manufacturers instead of building large, centralised plants. By working with Cell Agritech, the company will tap into regulatory-ready pilot and scale-up infrastructure, lowering capital requirements and reducing execution risk.

• Aleph’s hub-and-spoke setup pairs Singapore’s R&D and regulatory base with Malaysia’s larger, lower-cost manufacturing footprint, allowing a step-by-step expansion as regional demand grows.

• Aleph Farms’ flagship product, Petit Steak, is already cleared for sale in Israel and is under review in several other markets. The company is targeting restaurant launches in Israel and Singapore by 2027.

Source: Green Queen

BIO BUCKS

Funding, M&As, and grants

🐔 Orbem raised €55.5M Series B to expand its AI-based MRI inspection systems across poultry and produce markets

• The German company will use the funds to enter the US, expand its poultry offering, and add fruit and vegetable quality checks. Its Magnetic Resonance Imaging (MRI) scanners run in under a second and operate automatically. Orbem says it has already scanned more than 170 million eggs.

• Orbem’s flagship products, Genus Focus and Genus Scale, allow hatcheries to identify egg sex and fertilisation status without opening the egg, replacing inefficient or banned practices like male chick culling. These solutions also create new revenue streams and significantly reduce food loss.

• Since its founding, Orbem has scanned eggs, fruits, and seeds at high volume to build a large biological dataset. The company says this dataset will support future models that may carry insights from agriculture into non-invasive human health uses.

Investors: Innovation Industries, Supernova Invest, General Catalyst, 83NORTH, The Venture Collective, Possible Ventures, and several angel investors.

Source: EU-Startups

Thoughts 🤔:

Orbem’s story is a sign that internal quality data is becoming operational. If MRI can run at line speed, non-destructively, and without constant human tuning, it creates a new QC layer. That layer is routine internal phenotyping in hatcheries and potentially packhouses, going beyond surface proxies.

Once “inside quality” is measurable at scale, it can move from a private sorting edge to a tradable attribute. That could tighten specs, reduce disputes, and enable outcome-based contracts. But only if buyers accept scan-derived grades as auditable and stable across seasons, varieties, and facilities.

🧬 Barnwell Bio raised $6M to turn barn microbiomes into early outbreak signals

• Building from pandemic-era biosurveillance, Barnwell is deploying barn-level monitoring using metagenomic sequencing, starting with facilities in the Midwest and Southeast and is working with academic and industry partners.

• The New York-based startup doesn’t run one-off tests for a single pathogen. It samples the whole barn environment by collecting foot swabs that pick up faecal material and the broader microbial ecosystem. Those samples create a microbiome fingerprint for each facility, so producers can spot unusual spikes and track whether the barn’s microbial balance is shifting.

• The company turns results into risk scores and simple visuals that inform day-to-day decisions. Producers and veterinarians can step in sooner with targeted antibiotics, biosecurity upgrades, or nutritional and water-based treatment.

Investors: Twelve Below, Max Ventures, Dorm Room Fund, Banter Capital, Planeteer Capital, AgVentures Alliance, Daybreak Ventures, and Alumni Ventures.

Source: AgFunder

Thoughts 🤔:

Poultry barns are essentially closed, high-density ecosystems. If you can read that ecosystem through routine environmental sampling, you get a path to earlier, cheaper signal than waiting for sick birds or running one-off diagnostic panels.

Could the COVID playbook of population surveillance translate into farm-environment surveillance as a standard operating layer for animal ag? It only becomes real infrastructure if turnaround times fit day-to-day decision cycles and if outputs are tied to clear playbooks (what changes in biosecurity, treatment, or management when risk moves).

Metagenomics pushes the frame beyond “is pathogen X present?” toward “is this barn drifting from its normal microbial state?” That could enable systems-level health management beyond outbreaks, because you are tracking stability and stress over time.

🎨 Octarine Bio raised €5M to bring its precision-fermented pigments to market for food, textiles, and personal care

• The Danish startup has now raised €12.8M in its Series A. It plans to move from industrial validation into sales, while continuing work with existing partners. The funding will support PurePalette’s fully bio-based pigment line, produced through precision fermentation, with an initial market launch planned for 2026.

• This year, Octarine Bio is scaling three lead colours: PurePurple, PureGreen, and PureBlue, and doing formulation work with partners. The company has already produced 100kg of its first flagship pigment, and says this round will help bring more colours to the same stage.

• Octarine Bio says its pigments can match synthetic dyes on price even at early volumes, pointing to double-digit yields per litre, high colour strength, and stable performance. The process cuts environmental impact and offers an alternative as regulators and consumers scrutinise synthetic colourants more closely.

Investors: The Footprint Firm, Edaphon, Unconventional Ventures, DSM-Firmenich Ventures, Óskare Capital, and angel investors.

Source: Green Queen

🌾 Heritable Agriculture secured $4.98M grant from the Gates Foundation for AI-driven climate-resilient crops for smallholder farmers

• The funding will support the company’s “Joint AI-driven Smallholder Omics aNalytics” (JASON) initiative, which combines AI with genomic “omics” tools to accelerate crop improvement. The goal is to speed up gene discovery and shorten breeding cycles to bring resilient crop varieties to farmers more quickly.

• JASON helps smallholder farmers in low- and middle-income countries (LMICs), where rising heat, drought, and other climate pressures are increasing risks to food security. California-based Heritable integrates AI with genomic, remote-sensing, and historical crop data to identify traits that enhance stress tolerance and develop resilient germplasm.

• The effort comes as climate-driven yield losses intensify in smallholder systems, which produce up to 80% of food in LMICs. The initiative also aligns with the Gates Foundation’s push for digital innovation in agriculture. It could create partnership opportunities for seed companies, agribusinesses, and investors looking to strengthen climate-resilient supply chains.

Source: FoodBev

Thoughts 🤔:

The promise here is an engine that uses genomics, remote sensing, and other data to spot useful traits faster, then translate them into climate-adapted seed for smallholder systems.

That’s a big deal because weather volatility is accelerating, while new crop varieties still take years to develop and prove in the real world. If you can compress that cycle, it becomes a form of climate adaptation.

The risk is that prediction speed gets ahead of reality. Field proof still takes seasons, and real impact still depends on seed multiplication plus trusted local partners who can distribute seed and support adoption.

Will this produce farmer-ready lines that perform reliably across sites and years, and is there a credible path to getting those seeds into farmers’ hands at scale?

🍄 Maia Farms secured C$3.75M to expand specialty mushroom and mycelium ingredients focused on functional and nutritional health

• The Vancouver-based startup says it is shifting from “farm to pharmacy,” positioning its ingredients as part of a broader move toward health-led food innovation that puts public health at the centre of everyday eating.

• Maia Farms has developed two proprietary processes: Maia Form, a dry extrusion process that produces shelf-stable textured mushroom ingredients, and Maia Fresh, a fermentation-based process that yields high-moisture mycelium pulp. Together, they support up to 200,000 kg of capacity and have already powered 20+ product launches across North America.

• The company says its ingredients can compete on cost and performance. At a 4oz serving, they’re cheaper than beef and chicken mince, while still delivering functional benefits at low inclusion rates.

Investors: Active Impact Investments, NYA Ventures, Ag-West Bio, PIC Investment Group, and Deep Checks.

Source: Green Queen

BIO READS

Thought-provoking opinions

📈How the state’s ‘visible hand’ is driving China’s new agrifood investment playbook

• China’s agrifood tech scene is still moving fast despite the global slowdown. This is because state strategy actively builds the enabling infrastructure. High-standard farmland serves as a large-scale innovation testbed, and biomanufacturing clusters help reduce scale-up friction for synthetic biology platforms.

• Matilda’s core claim is that “what’s investable” in China is shifting to strategic fit. Companies that solve national priority chokepoints get policy support, infrastructure, and capital pathways that make commercialisation and scale more feasible than in many other markets.

• She also argues liquidity is returning through a mix of reopened public markets, stronger M&A channels, and government-backed funds that can finance capital-intensive growth stages and provide exits. This creates a tighter public-private loop where the state supplies patient capital and private GPs execute.

Source: Matilda Ho (AgFunder)

🚜 Why autonomous machinery in agriculture hasn’t scaled (yet)

• Autonomous farm machinery has made real technical progress, moving from early retrofit experiments to hundreds of deployed units, but it is still tiny compared to the global equipment fleet.

• Patrick argues that the usual explanations (regulation, farm complexity, rural connectivity) are incomplete. The more immediate blockers are safety proof, unclear liability, farmer trust, and systems that are not yet robust or versatile enough to remove labour rather than shift it.

• Even as ROI improves through falling hardware costs and models like subscriptions or robotics-as-a-service, adoption will remain slow until autonomy works predictably across tasks and farm sizes, and until the surrounding standards and support ecosystem catch up.

Source: Patrick Honcoop

🤢 Where disgust meets distrust: The viral psychology behind the “Ultra-Processed Food” narrative

• Dr. Sophie argues the “ultra-processed food” story spreads because it plays on our food neophobia. It treats processing and unfamiliar ingredients as contamination, which cues disgust and fear.

• She says this lumps nutritionally weak junk foods together with a wide range of often beneficial staples and newer climate-friendly innovations like plant-based, upcycled, and fermentation-derived foods, which could be important for food security and sustainability.

• To blunt the disgust-and-distrust loop, the piece suggests changing how new foods are sold and talked about. That could mean drawing clear lines between categories, presenting novelty as craft or status, and building familiarity through formats people already trust.

Source: Dr. Sophie Attwood (Green Queen)

If you found value in this newsletter, consider sharing it with a friend who might benefit! Or, if someone forwarded this to you, consider subscribing.

Share

Subscribe now

Thank you, and have a great day!

Disclaimer: The views and opinions expressed in this newsletter are my own and do not necessarily reflect those of my employer, affiliates, or any organisations I am associated with.

This week’s issue spans the full stack, from consumer-facing product launches to deep infrastructure bets. On the food side, I look at how protein is “menu-modified” into everyday routines in India and what that says about foodservice as a scaling channel for new ingredients.

In gut health, Bactolife’s €30M round points to a potential shift from reshaping the microbiome to controlling its outputs, with implications for how claims, dosing, and trials are designed.

On the farm, several stories converge on the same question: what actually works at scale. RNAi-based crop protection is moving past target novelty toward delivery, stability, and cost per hectare. AI-driven trait discovery is reframing where defensibility really lies as gene-editing tools commoditise. Autonomous mustering drones highlight a different agtech wedge altogether, shifting from per-animal hardware to ranch-level capacity economics.

I also cover a notable jump in South Korea’s public R&D spend, a strategic biocontrol acquisition by BASF, and why AI agents, not just GLP-1s, could represent a more serious structural threat to Big Food.

Let’s get into it!

BIO BUZZ

Products, partnerships, and regulations

🇮🇳 Starbucks India launches a new line of cold foams made with SuperYou’s yeast protein

• Starbucks India has launched protein cold foams across all 500 stores, tapping into India’s growing demand for high-protein functional foods, with each serving adding 11-18g of protein for ₹50 (~$0.55).

• It also has added probiotics for gut wellness. India is the first market outside North America to get Starbucks’ protein cold foams, and the first to offer them as a non-dairy option. SuperYou’s protein powder is produced from brewer’s yeast grown on molasses in bioreactors and later purified.

• The lineup includes sugar-free vanilla, chocolate, and banana. SuperYou says the powder provides 24–27g of protein per 36g serving, includes all nine amino acids, and has a PDCAAS score of 1.0.

Source: Green Queen

Thoughts 🤔:

Protein is getting “menu-modified” into everyday routines. A ₹50 add-on delivering ~18g reframes protein from a product category into a default upsell, like extra espresso or oat milk. The question is whether this survives beyond novelty: repeat orders will tell us if proteinization is moving from fitness to habit.

It also highlights why foodservice can often scale ‘newer’ proteins faster than CPG shelves. Chains bring distribution and trust, but they are brutal filters on functionality. Foam stability, taste masking, speed-of-service, and supply reliability matter more than the protein’s origin story.

🇬🇧🇳🇱 Moa Technology and Certis Belchim partner to accelerate next-generation herbicide Moa Amplifier solutions for global agriculture

• Moa will bring its discovery engine, including more than 80 newly identified herbicidal modes of action, over the past three years, while Certis Belchim will handle development, registration, and commercialisation.

• The companies plan to co-develop a new Moa Amplifier molecule designed to work with a specific active ingredient. This is intended to help farmers protect yields more safely, sustainably, and effectively than current approaches.

• Moa Amplifiers don’t kill weeds on their own, but they can make herbicide programs work with lower doses or concentrations. Using the proprietary Moa Gamma platform, Moa has already identified hundreds of these amplifier candidates and is now advancing them with commercial partners.

Source: AgriBusiness Global

Thoughts 🤔:

Globally, resistance has been documented in 273 weed species across 21 of 31 known herbicide sites of action. That base rate is why ‘incremental’ innovations can still matter, if they reliably reduce dose or improve control across real-world variance.

Incumbents are starting to place more bets earlier on platform-derived efficacy enhancers because discovering and commercialising a truly new herbicide mode of action still takes a long time, and costs a lot, while resistance continues to compound.

🇧🇷 Vitalforce and Bsafe Biotech Inc to co-develop an RNAi-based bioinsecticide for fall armyworm

• RNA interference (RNAi) works by using double-stranded RNA to trigger degradation of a target organism’s messenger RNA, preventing production of vital proteins. Since the mechanism is sequence-specific, it can deliver narrower-spectrum control with fewer non-target effects than conventional broad-spectrum insecticides.

• The partnership targets fall armyworm, one of the most economically damaging pests in tropical agriculture, especially in Brazil’s commodity systems. Embrapa-documented losses of 30% in soybeans and 40% in corn under heavy infestation underscore why fall armyworm’s biology and resistance profile make it a top priority.

• The program is structured around an 18-month R&D cycle followed by regulatory preparations, with milestones split into 9 months of lab efficacy work and 9 months of field validation. If field performance is strong, they plan to compile the safety/efficacy dossier and pursue commercialisation in late 2027 to early 2028.

Source: AgroPages

Thoughts 🤔:

Brazil is the stress-test market for next-gen insect control. The scale and year-round pest pressure speed up resistance and expose weak links in agronomy, logistics, and price.

As Bt performance slips, growers are demanding new modes of action. If RNAi insecticides can deliver repeatable field control across regions, stay stable in heat and humidity, hit a workable cost per hectare, and clear Brazil’s three-agency safety review, they’ll be de-risked for other high-pressure geographies.

🇬🇧 🇺🇸 Wild Bioscience and The Traits Company team up to speed development of new soybean varieties

• UK-based Wild Bioscience is partnering with The Traits Company, a North Carolina trait discovery specialist backed by global seed leader GDM, validating demand for Wild Bio’s AI-based trait design system in a crowded crop innovation market.

• Wild Bio uses machine learning and evolutionary biology to design traits for resilience and efficiency, cutting development timelines. The Traits Company will pair that design work with its soybean development pipeline to move candidates through testing and into market-ready varieties at scale.

• Backed by GDM’s global reach, the partnership will focus on next-generation soybeans aimed at resilience, nutrition, and climate adaptability.

Source: AgTechNavigator

🇦🇺🇬🇧 CSIRO and the University of Leeds are developing an open-access AI tool to convert food waste into fermentation-based proteins

• The two-year project is funded by a $2M grant from the Bezos Earth Fund. It will use farm and food sidestreams, including damaged or unharvested vegetable crops, grain leftovers such as canola meal, rice bran, and brewer’s spent grain, and byproducts from cheese production.

• Around one-third of the world’s food is wasted, creating about 10% of global emissions and roughly $1 trillion in economic losses.

• The team plans to use AI-tuned biomass fermentation to convert these sidestreams into microbial protein powders that can compete with conventional proteins on cost and performance for people and animals.

Source: Green Queen

BIO BUCKS

Funding, M&As, and grants

🇩🇰 Bactolife raised €30M to launch its gut-health-boosting binding proteins and start human study programme

• The Danish startup is building a new functional gut-health ingredient category using binding proteins inspired by IgG fragments found in camelid milk (camels, alpacas, llamas), produced using precision fermentation.

• Binding proteins bind specific harmful metabolites (toxins) so they pass through the GI tract without disrupting the gut microbiome. Unlike antibiotics, they don’t kill gut bacteria, don’t penetrate the gut barrier, and don’t activate the immune system. They also work at low daily intake, with a fast onset and observable impact soon after ingestion.

• The new funds will support a human study programme spanning the US, Asia, and low- and middle-income countries, alongside investments in production and supply capacity. Commercially, the rollout begins in the US before expanding to the EU and Asia, while also aiming to improve access for women and children in lower-income markets.

Investors: Cross Border Impact Ventures, EIFO, Novo Holdings, and Athos.

Source: Green Queen

Thoughts 🤔:

Bactolife is a nice marker for a potential shift in gut-health ingredients. It’s a move away from “feeding the microbiome” and toward “controlling outputs” (metabolites) with binders. The idea isn’t new in principle (gut sequestration has plenty of precedents), but the bet here is higher selectivity and a cleaner mechanism-to-claim pathway.

I’m curious to see: the trial endpoints that move biomarkers, disclosed effective dose and cost per serving, no meaningful off-target binding (nutrients/meds), and early repeat-order rates post-launch.

🇬🇧 Biographica raised a £7M seed round to scale its AI-driven crop trait discovery platform

• Traditional crop development suffers from a <1% hit rate, forcing companies to test thousands of edits to find a single success. The London-based startup’s AI-driven approach streamlines this by identifying high-value targets 12x faster, significantly reducing the time and risk involved in innovation.

• Unlike traditional GWAS methods that only show correlation, Biographica uses knowledge graphs and foundation models to understand the mechanistic “why” behind genetic traits. This “lab-in-the-loop” system allows the AI to self-improve by constantly integrating experimental feedback into its predictive algorithms.

• The company trained its models on large public and self-generated datasets. This strategy has already secured trust and active R&D projects with two of the top-five global seed firms.

Investors: Faber VC, SuperSeed, Cardumen Capital, The Helm, and Chalfen Ventures and Entrepreneurs First.

Source: AgFunder

Thoughts 🤔:

As gene editing is increasingly commoditized, the scarce capability is upstream. The hard part is identifying causal targets and choosing edit strategies that survive the messy reality of pleiotropy (single gene influencing multiple traits) and environment-by-genotype effects.

A platform can reliably reduce the number of edit events, lines, and seasons needed to land a usable trait, could shift trait-development unit economics more than incremental improvements in CRISPR tools.

🇩🇪 SenseUP Biosciences raised €3M seed funding to expand its dsRNA-based biopesticide portfolio that addresses cost, stability, and use case challenges

• The German startup is building RNA interference (RNAi) sprays that use double-stranded RNA (dsRNA) to switch off specific genes in target pests. The goal is targeted control without the broad collateral damage linked to many chemical pesticides.

• The patented Corynebacterium-based platform tackles some of RNAi’s biggest hurdles. It keeps dsRNA stable for more than 18 months at room temperature, reduces manufacturing costs through high-throughput strain screening, and supports multi-pest control in a single formulation.

• The company says early field trials are promising, and it has about 15 products in development. Larger trial readouts are expected in 2026. SenseUP plans to file regulatory submissions with large partners, starting with the US and South America, with Europe also on the roadmap.

Investors: Capnamic, Simon Capital, Rockstart, CHECK24 Impact and HBG Ventures.

Source: AgFunder

Thoughts 🤔:

With commercialisation in mind, the differentiator in RNAi crop protection is shifting from target novelty to delivery and unit economics. The key is consistent field performance at a cost per hectare farmers will tolerate, in a product that fits existing spray workflows. Shelf-stable biology is becoming table stakes because logistics friction is a quiet adoption killer.

Encapsulating dsRNA inside microbial cell walls reframes RNAi as a fermentation and formulation problem (titer, downstream simplicity, shelf life), not just RNA design. It would be interesting to see the cost per acre at commercial dose, and how regulators treat multi-target and cell-based formulations.

🇦🇺 GrazeMate raised $1.2M in pre-seed funding to launch fully autonomous cattle-mustering drones run from a mobile app

• Ranchers can send a drone out from their phone to move cattle, cutting hours spent on horses, bikes, and other vehicles. This directly tackles the “hair-on-fire” problem of skilled labour being scarce, expensive, and time-intensive.

• Early pilots in Queensland and New South Wales cover 1.7 million acres and are mustering thousands of cattle each week. The drones use reinforcement learning to respond to cattle in real time, backing off when animals show signs of stress. The system can also handle monitoring, and a second-generation beta estimates cattle weight and dry matter availability.

• Instead of selling hardware, GrazeMate leases drones and charges a monthly fee based on ranch size and cattle numbers, aiming to undercut current mustering costs. The value proposition improves as cattle are moved more frequently, supporting rotational grazing and regenerative practices that lead to better grass and soil over time.

Investors: Y Combinator, Antler and NextGen Ventures.

Source: AgFunder

Thoughts 🤔:

If GrazeMate can prove that one off-the-shelf drone and a base station can take a big chunk of mustering hours off large ranches, it’s a shift from per-head hardware economics (smart collars) to ranch-level capacity economics (shared autonomy).

But that only holds if the system keeps working in real terrain and weather, supply risk (DJI) don’t force a big jump in CAPEX or OPEX, and customers keep renewing even with seasonal use.

The key is cost lumpiness. Collars scale roughly linearly with herd size because every additional animal adds hardware and subscription cost. Drone mustering looks like a capacity service because once a ranch has enough drone and base coverage, adding animals shouldn’t cost much.

🇩🇪🇺🇸 BASF Agricultural Solutions acquires AgBiTech from Paine Schwartz Partners to strengthen its biocontrol capabilities

• Founded in 2000, Texas-based AgBiTech develops insect control products using nucleopolyhedrovirus technology for corn, cotton, soybean, and specialty crops. Its solutions are already commercialised across Australia, Brazil, and the U.S., serving major row crops like corn, cotton, and soybeans.

• The acquisition (amount undisclosed ) is expected to close in the first half of 2026, pending regulatory approval. It will complement BASF’s existing biological portfolio aimed at disease management, residue reduction, stress tolerance, and resistance management.

• Brazil is a major strategic opportunity for BASF due to its rapidly expanding biologicals market and recurring pest pressures. AgBiTech’s proven solutions against lepidoptera caterpillars address a critical challenge for Brazilian farmers.

Source: AgTechNavigator

🇰🇷 South Korea’s agriculture ministry plans to invest KRW 234.8B ($162.2M) in agricultural R&D in 2026, up 17% from last year

• Climate resilience and risk management received the largest allocation, with KRW 73.2B allocated to address climate change impacts, emerging diseases, and agricultural disasters.

• To modernise farming systems, KRW 51B is earmarked for smart agriculture, including AI, robotics, and drone technologies. These investments are intended to raise productivity, optimise resource use, and support large-scale precision farming.

• Beyond production, funding is spread across biotech (KRW 33.5B), commercialisation and talent development (KRW 38.2B), and future food and K-Food innovation (KRW 38.9B) to strengthen global competitiveness.

Source: AgTechNavigator

BIO READS

Thought-provoking opinions

⚖️ Agriculture’s great repricing goes beyond the downcycle

• Mark’s core claim is that agriculture is not in a normal downcycle. Markets are pricing it as a riskier system, so the cost of capital rises as moats fade faster and regulation, litigation, and geopolitics move to the centre.

• He points to surface signals across the stack: FMC down ~80% in market cap from end-2023 to end-2025, agrifoodtech VC dropping, and 25,000+ strategic role cuts since 2023. His take is that the industry is starting to look like pharma: AI makes discovery cheaper, but development, registration, manufacturing, and adoption stay the bottlenecks.

• So the money shifts to choke points that compound: formulation and manufacturing, IP platforms, distribution and advisory networks, finance tied to inputs and offtake, and data plus verification. Consolidation becomes defensive, and the next cycle favours long-horizon operators over pure discovery engines.

Source: Mark S. Brooks

🤖 The end of impulse: Why AI agents are Big Food’s existential threat

• Antony argues that Big Food is mispriced if investors treat GLP-1s as a modest 1-3% volume headwind. The bigger risk is a shift in how people buy: AI agents become the shopping interface, turning grocery into a planned reorder and squeezing out impulse purchases.

• He adds other shocks that could stack on top: stricter dietary guidance aimed at ultra-processed foods, rising litigation and liability risk that is starting to look like the tobacco playbook, retailers moving shelf space toward healthier private-label and perimeter categories, and insurers pricing diet into premiums.

• He ends with a capital rotation point: the spending won’t vanish, but it may move to products and supply chains that can prove “quality” with data. This pushes the advantage upstream into biology, provenance, and measurable nutrient density.

Source: Antony Yousefian

🧫 China is coming for agricultural machinery: Are Western OEMs ready?

• Chinese agricultural machinery makers are starting to follow the automotive EV playbook: move from “low-cost manufacturing” into credible, feature-rich products, then export aggressively, with Agritechnica 2025 as a visible inflection point.

• Patrick maps the challengers (e.g., YTO, Foton Lovol, Zoomlion) and the conditions for a real Western breakthrough: step-change quality and compliance, credible dealer and service networks, trust and brand building (including acquisitions), and financing and residual-value support that makes total cost of ownership workable for farmers.

• Western OEMs should take note. As dealers consolidate and tech becomes less distinct, incumbents lose protection. The response is to move faster on product, tighten customer value, and treat Chinese entrants as a real medium-term threat.

Source: Patrick Honcoop

EAR FOOD

Podcast episode of the week

🎙 David Friedberg: ‘Transformative ag companies can be built in any macro environment to drive outlier VC returns’

• “Farm impact” has to be a loud signal. Friedberg’s bar is: if gains don’t clearly beat field noise, the market won’t pay. In ag, uncertainty and seasonality make it hard for growers to trust marginal gains, so the value prop needs to be both measurable and repeatable fast.

• Ohalo’s entry point is turning vegetative crops into true-seed markets, which changes farm economics. The potato example: instead of hauling ~4 tons of planting material per acre (plus storage, fungicide, diesel, and labour), you plant small, uniform true seed and aim for both lower costs and higher yields.

• Pricing and adoption follow the usual “share the upside” rule: charge about one-third of the added profit you create. He frames this as the Monsanto-style seed model and analogous to enterprise software pricing (capture a slice of created value,).

• Agtech is structurally slower to scale than typical software because the “iteration loop” is annual. Capital efficiency is constrained by once-a-year selling windows and waiting a full season for results before customers renew/expand and before product teams can iterate. That cadence caps how much capital the sector can productively absorb.

• He thinks the sector may be normalising after a long bubble, not just in a temporary trough, and that this will be painful for undifferentiated VC funds. But power-law reality doesn’t change: truly transformative companies can still be built in any macro environment, and most VC returns depend on rare outliers rather than “the category.”

Guest: David Friedberg, CEO and co-founder of Ohalo Genetics

Podcast: The Modern Acre

If you found value in this newsletter, consider sharing it with a friend who might benefit! Or, if someone forwarded this to you, consider subscribing.

Share

Subscribe now

Thank you, and have a great day!

Disclaimer: The views and opinions expressed in this newsletter are my own and do not necessarily reflect those of my employer, affiliates, or any organisations I am associated with.

Thanks for being here.

For Issue #132 of Better Bioeconomy, I spoke with Cyril Ebersweiler, General Partner at SOSV, a global deep tech venture firm, and the founder and co-Managing Director of HAX, SOSV’s hard tech startup development program.

Cyril backs deep tech across hardware, robotics, biotech, and health, and has spent the past decade helping founders turn early prototypes into real companies. Last year, he launched RETVRN, a program designed to help early-stage deep tech and hardware startups prepare for exits.

I start with what RETVRN’s State of Exits 2025 data says about liquidity right now. Then, move into the exit mistake founders repeat, the traits that predict strong outcomes, when to build versus sell, what exit-readiness looks like, what buyers hammer in diligence, and which acquirer types are getting more active post-2022.

Let’s jump in!

Why exit planning matters

When the headlines are all unicorns and IPOs, it’s easy to treat acquisitions as the “backup plan.” But for most venture portfolios, M&A is the plan.

After 2022, that became harder to ignore. Valuations fell, risk appetite dropped, and buyers got pickier about what they’ll pay for. In that market, “we’ll think about an exit later” turns into “we’re scrambling when an offer shows up” or “we’re scrambling when the runway runs out.” Exit planning prevents both. It forces you to align early on what a great outcome looks like for your company, your team, and your cap table, before urgency limits your options.

An exit isn’t just a number on the paper. It’s the result of years of choices: ownership structure, IP paperwork, how clean the books are, how expectations are managed, and whether outsiders can understand what you’ve built. A sloppy process can leave founders with weak bargaining power, investors pulling in different directions, and a cap table that slows everything down.

A well-timed sale can do the opposite: put a technology inside a big distribution engine and get it into the world faster. As Cyril told me, “An exit to a strategic buyer is only the beginning of a new adventure, never forget that!”

Exits also keep the startup ecosystem alive. When founders and early teams get liquidity, some of that capital turns into new startups, angel checks, and operator talent recycled back into the next wave. When investors return capital to LPs, new funds get raised, risk budgets reset, and more startups get funded.

Before we jump into the conversation, here’s what RETVRN’s latest report says about the market.

What the exit market looks like right now

The report paints a stark picture of how value is captured. In 2024, 71% of exit dollars came from secondaries (investors selling shares to other investors), not IPOs or M&A, and 96% of M&A deals were under US$500M, and 70% were under US$100M.

There’s also a massive liquidity gap between the capital flowing into startups and the value flowing out. RETVRN estimates roughly US$180B went into deep tech in 2024, while exit value was about US$89B, implying a funding‑to‑exit ratio of roughly two to one.

The shape of the exit market is changing, too. Global M&A volume fell about 9% in the first half of 2025, while average deal value rose 15%. This reflects a “flight to quality” in which buyers are doing fewer but larger deals.

Meanwhile, secondaries have also grown fast. The report puts them at 28% of exit volume (up from ~10% in 2020), and rising from 10% to 48% of overall exit market share.

Deal structures themselves are evolving. The report notes that 73% of transactions include earn‑out provisions (future payments tied to performance), and roughly 42 % of total consideration is contingent. Earn‑outs now average ~3.2 years and stretch across product milestones, revenue targets and integration milestones.

Timing is another surprise. More than 60% of acquisitions happen at or before Series A, and 47% involve seed-stage companies. The highest concentration sits below US$10-15M raised: 34% of deals happen there, and 46% of all deals happen within the first five years

Sector differences matter. RETVRN reports AI/ML companies reach median valuations ~3.2 times higher than the overall deep tech median. SaaS multiples fell from ~15 times revenue in 2021 to ~7 times. For biotech, the top quartile reaches 8-10 times the median. Time-to-exit differs too: AI often sells in 3-5 years, SaaS ~5.2 years, hardware 6-8, and biotech 11 years.

Finally, the buyer mix is also broadening. The report says 85% of strategic acquirers cite AI/ML as a top priority, and 72% cite cybersecurity. Sovereign wealth funds are showing up more often: direct investment into deep tech hit US$47B in 2024, and sovereign-led deals are up 3.2 times since 2020.

The most common exit mistakes

We kicked off our conversation with a simple question: What’s the single most common exit mistake you see? Cyril’s answer: “waiting too long to think about the exit.” Too many founders obsess over product and fundraising but neglect the downstream implications of how, when and to whom they might eventually sell.

“Founders often wake up at the last minute and realise they don’t know who their likely buyers are, or that their cap table is a mess. By then, it’s hard to clean up. You should start with an exit in mind and back‑track from there.”

In practice, that means sketching a buyer map early on. Who would benefit most from owning your IP? Which business unit wins if your tech exists inside their product line? What would integration look like? If you identify plausible acquirers in the first 12-18 months and track what they buy, you can make choices today that fit what they pay for later.

Another common mistake is failing to maintain a clean cap table. When investors, advisors and early employees accumulate layers of debt, warrants or SAFE notes with different terms, the eventual exit can get bogged down by renegotiations. A streamlined cap table with clear share classes, clear vesting, and no hidden obligations makes diligence smoother and increases the take‑home proceeds for everyone involved.

Finally, Cyril warned against neglecting governance. “Misaligned boards can kill an exit,” he said. If one investor won’t sell without a 10 times return, or co-founders no longer want the same future, even a fair offer becomes impossible to close. Talk through exit scenarios early, define decision rights, and agree on downside protections while everyone still has time to think.

Subscribe now

What strong exits have in common

Cyril has seen dozens of exits across hardware and biotech through SOSV’s accelerators. So, what early decisions most consistently set up a company for a strong M&A outcome, even if it never reaches unicorn scale?

First, match the technology readiness level (TRL) to the capital you raise. Raising only what you need to reach proof-of-concept and early revenue (TRL 4-6) keeps expectations grounded. Huge rounds on speculative tech can price you out of what strategics are willing to pay.

Second, have a clear buyer logic. The most successful exits have a clear narrative connecting the startup’s capability to the acquirer’s strategy, which could be a technology bolt‑on, a market entry or a talent acquisition.

Third is early customer traction. Even in deep tech, a handful of paying customers changes the conversation. It shows the problem is real and gives buyers data they can underwrite.

Finally, relationships compound. People underestimate the time horizon of relationships. Introductions made in year one can pay off in year five. Joining pilot programs, co‑presenting at conferences and collaborating on open innovation initiatives are ways to get on the radar of potential buyers.

Navigating the trade‑off of building more or selling now

One of the hardest decisions founders face is whether to double down on building, i.e., raising more capital, expanding the product and aim for a bigger exit later, or to sell early.

Cyril suggested looking at three axes: technology maturity, market timing, and capital intensity. If you’re at TRL 4-5 and can reach TRL 7 with a reasonable amount of capital, staying independent may pay off because each milestone reduces buyer risk and can raise price. If the next step takes tens of millions and years of regulatory work, and buyers are already circling, selling could make more sense.

Another consideration is macroeconomic cycles. In the post-2022 market, deal volume fell, but strong companies still get paid for. If your investors are still anchored to 2021, waiting may not change much. If policy or supply-chain shifts are pushing on-shoring or local manufacturing, staying in the fight longer could capture more value.

Cyril suggested that founders ask themselves what their business will look like in three years if they do not sell, mapping out the capital required, milestones achieved and burn rate. Then map how the buyer set might change whether potential acquirers will be larger, hungrier or more regulated and whether new entrants such as sovereign funds or private equity will join the mix.

They must also assess whether they have enough runway and board alignment to make that journey, because if investors have limited reserves or different return horizons, pushing for an extra year may not be feasible.

Diligence red flags and how to avoid them

When a startup enters a sale process, the diligence phase can make or break a deal, and it’s rarely small. Cyril said even modest acquisitions can trigger 500+ diligence requests.

Common points of failure tend to show up in the same places. On the operations side, messy financials and a data room built at the last minute slow everything down. Buyers expect numbers accurate within ±10-15% and a process that can close fast.

On the legal and IP side, unclear IP assignment, contract loopholes, restrictive licensing, and filings that don’t match the story raise risk. Cap table surprises, such as odd clauses, side deals, and unclear proceeds flow, can derail terms. Strategy breaks when founders can’t explain why this buyer should buy now. And deals stall when leadership and governance aren’t aligned.

Treat diligence as ongoing work. Keep the data room current, track covenant compliance, and stay on top of IP filings.

Who buys deep tech now and why

With buyers more selective after 2022, Cyril sees two groups showing up more often: mandate‑driven strategics and sovereign and national‑security buyers.

Mandate‑driven strategics are big companies with clear goals in AI/ML, cybersecurity, data infrastructure, and “picks and shovels” categories that are buying to move faster. It’s not only Big Tech, but automotive and industrial companies are also buying robotics and AI to ship products and cut costs.

Sovereign and national‑security buyers include governments and sovereign wealth funds, which are playing a larger role in deep tech M&A. This group is often motivated by national self‑sufficiency or strategic advantage and comes with conditions like on‑shore manufacturing or technology transfer.

At the same time, private equity is increasingly active in deep tech, typically looking for profitable niche players where they can drive operational efficiencies and roll up adjacent capabilities. This group is particularly relevant for hardware companies that have reached breakeven but need scale to compete.

The net effect is a more fragmented but opportunity‑rich buyer landscape, and founders should broaden their horizon beyond the obvious names and consider corporate VCs, sovereign investors and PE funds.

What you can do today

Here is how you can operationalise your exit strategy starting today:

• Build a buyer map early. In the first 12-18 months, list 5-10 strategic acquirers who would benefit most from your IP or market position. Don’t just list “Google” or “Bayer”, try to identify the specific business unit whose pain point your technology solves.

• Build relationships ahead of need. Aim for active corporate development or strategic partnership threads 18-24 months before you want the option to sell. Pilots and JDAs are a practical way to prove value.

• Keep your diligence basics clean. Treat your data room as a living document, not a project for the final hour. Ensure all IP assignments are signed, employee contracts are clean, and financials are accurate within a ±10-15% variance.

• Check the capital treadmill before the next round. Map your TRL against the cash required to reach the next milestone. If the cost to reach TRL 7 forces major dilution that prices you out of 96% of the M&A market, weigh an earlier sale.

• Make the bolt-on case easy. Your story should answer: where does this plug into the buyer’s product, and why does it pay off?

• Get the board on the same page about the number. Talk about exit scenarios early. Ensure everyone is aligned on what a “win” looks like so you can act decisively when an offer arrives.

Want more investor perspectives like this? You might like my agrifood investor interview series, where I interview the investors backing the technologies reshaping food and ag. Recent conversations include PeakBridge, Big Idea Ventures, Ajinomoto Group Ventures, and more.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

For Issue #131 of Better Bioeconomy, I’m digging into two recent sources I read over the holidays, one from McKinsey and one from BCG. Both look at a question I think about: why have publicly traded agrifood companies delivered relatively low returns, and what do the consistent winners do differently?

What follows is my attempt to stitch these two reports together: McKinsey’s long-run diagnosis and “big moves,” and BCG’s five-year stress test of what created value from 2020 to 2024.

Let’s dig in! 🍽️

Agrifood is one of those sectors that seem “obviously investable.”

People have to eat. Demand grows with population and income. And when you zoom out far enough, the world is running a multi-decade experiment in reengineering how we produce calories, protein, and inputs under tighter constraints: land, water, labour, and carbon.

However, in public markets, agrifood has not been a consistent creator of shareholder value.

McKinsey’s analysis of 134 publicly traded agrifood companies finds that agriculture delivered an average TSR (Total Shareholder Returns) CAGR of about 6% from 2010 to 2025, trailing the S&P 500 since 2010. TSR is the total return an investor earns from owning a stock: share price gains plus dividends. CAGR (Compound Annual Growth Rate) means the average annual rate over a period.

BCG (Boston Consulting Group)’s 2025 Agribusiness Value Creators report reaches a similar conclusion. Among 38 publicly traded agribusiness companies, the median TSR from 2020 to 2024 was 4%, and agribusiness ranked near the bottom of BCG’s sector set, beating only 4 of 36.

Both reports land on the same idea. Agrifood is cyclical, but cycles do not fully explain weak returns. A small group has managed to outperform again and again. The gap comes down to a few practical things: how leaders invest capital, how tightly they manage cash tied up in the business, how well they defend margins, and whether they sell something customers will pay more for.

In other words, the winners are not “riding the commodity cycle.” They are running a different operating system.

What the McKinsey report says

McKinsey frames agriculture’s performance problem as structural.

Volatility is part of the job. Over the past 15 years, agriculture has moved through several distinct phases: a biofuels-driven expansion (2009 to 2013), a return to trend (2013 to 2019), an inflation reset (2019 to 2022), and another return to trend (2022 to 2025).

The bigger issue is that many agrifood companies have not turned that volatility into durable value. McKinsey points to a few recurring drags.

1. Innovation intensity has weakened

Agrifood often underinvests in innovation relative to the complexity of the problems it faces. Median R&D spend has been flat at about 1% of revenue since 2010. Adjusting for inflation, real R&D investment has contracted over time.

In many subsectors, innovation is how you stand out (differentiate). Standing out helps you protect margins. Healthy margins give you room to invest during downturns.

A low-R&D loop tends to create a self-reinforcing trap: underinvest, compete more on price, see weaker margins, then underinvest again.

2. Digital adoption is uneven and often fails to translate into productivity

Digital tools have spread, but not evenly, and not always with the productivity gains people hoped for.

In a 2024 McKinsey survey, about half of US farmers reported using precision agriculture hardware. Adoption was around 30% in Brazil and the EU, and around 5% in India. Adoption of remote sensing, farm management software, and automation was lower.

This matters for two reasons. First, it limits near-term productivity gains across the system. Second, it means many companies are selling tools into markets where adoption is slow and willingness to pay is limited. The value has to be specific and proven.

3. Capital productivity has gotten worse

Agrifood is asset-heavy. It runs on infrastructure, equipment, storage, processing capacity, and distribution networks.

In asset-heavy businesses, small changes in how well assets are used can swing results. McKinsey notes that fixed capital is large relative to the value of goods produced, and fixed asset turns have not improved much since 2010.

4. Working capital has become a drag

McKinsey estimates the sector’s median cash conversion cycle has lengthened by more than 32 days since 2010. The cash conversion cycle is how long cash is tied up after you pay suppliers: you buy inputs, hold inventory, sell, then wait to get paid.

Inventory turnover (how quickly inventory gets sold and replaced) has also dropped. Together, this has pushed the sector to carry more than $60 billion in additional working capital versus 2010.

Working capital (cash tied up in day-to-day operations, mainly inventory and unpaid customer invoices) often looks like accounting. In cyclical industries, it acts like strategy. When too much cash is stuck in inventory and unpaid invoices, you lose flexibility. And flexibility matters most in downturns, when assets and capabilities get cheaper.

What winners do differently

Despite volatility in commodity prices, the common link among leaders is superior ROIC (Return on Invested Capital), driven by greater capital efficiency. ROIC is a measure of how much profit a company generates for each dollar it invests in the business (factories, equipment, inventory, acquisitions). Capital efficiency is about getting more output and profit from the same set of assets, while keeping less cash stuck in inventory.

McKinsey gives examples across four subsectors:

Crop protection and biologicals leaders narrowed offerings toward IP-rich (patents, proprietary formulations, protected traits) segments (patent-protected or proprietary blends), shifted fixed costs to variable costs (including contract manufacturing), and tightened channels and trade spend.

Equipment OEM (original equipment manufacturer) leaders moved to build-to-demand models, reassessed dealers based on inventory and performance, and developed software-enhanced machines priced for farmer outcomes.

Fertiliser leaders reduced conversion costs, offered premium products (multinutrient blends, micronutrients with field-proven nutrient-use efficiency claims), and exited assets requiring sustained heavy CAPEX (large-scale long-term spending such as building plants, buying equipment, or expanding capacity).

Grain origination leaders increased capital turns, consolidated underutilised sites, rebalanced toward advantaged logistics corridors, built digital grower platforms to lower cost to serve, and decommodified outputs (traceable ingredients, lower-carbon-intensity options).

The five “big moves” that show up among outperformers

1. Programmatic M&A

This is not “do a big deal once in a while.” It’s building a repeatable machine for finding deals, evaluating them, integrating them, and learning from them.

McKinsey finds that for leading players, no single deal exceeds 30% of the company’s market value, while total deals add up to about 60% of market value over a decade.

The logic is risk control. One giant acquisition depends on timing and integration going perfectly. A steady cadence spreads risk and builds skill. It also lets companies buy capabilities when prices are lower.

2. Dynamic resource reallocation

Most companies say they allocate capital rationally. Many do not. In practice, many allocate capital by protecting legacy footprints, defending historical priorities, and avoiding politically painful exits.

McKinsey claims leaders move ~50% of their CAPEX across businesses over a decade. That suggests they keep reassessing where returns are best, and they move money accordingly.

3. Big investments (especially in downturns)

Data suggest that leaders in the top 20% of the industry spend approximately 1.7 times the median capital spending on sales.

“Big investments” mainly mean two things. It means being willing to invest in capabilities that change the long-term position of the business, such as R&D, capacity, and new business building. It also means acting when the cycle makes those investments cheaper.

This is where working capital discipline matters. It gives you the cash to invest when others have to pull back.

4. Productivity leadership

McKinsey frames productivity as something leaders operationalise. They point to levers like consolidation, automation, exiting weak units, improving sales productivity, and cutting overhead costs such as sales and administration.

The key is consistency. In cyclical businesses, productivity is what keeps the economics from falling apart when prices drop.

5. Product differentiation

This is often the most difficult step, because it can require letting go of what used to work.

McKinsey argues the best performers expand gross margin faster than peers by selling more differentiated offerings. That usually requires a clear view of where the market is going, and sometimes means improving or replacing older products before competitors force you to.

Differentiation is not branding. It is a real improvement in what the customer gets and what they are willing to pay for.

Agrifood companies can struggle here because many buyers are trained to view inputs as commodities. Changing that mindset often takes proof, service, and business models that price around outcomes.

Subscribe now

What the BCG report says

McKinsey looks across 15 years. BCG looks closely at the recent five-year stretch.

BCG calls 2020 to 2024 a “mixed bag.” Some companies benefited early from record-high commodity prices, but growth did not always turn into strong shareholder returns.

The headline number: median TSR across the 38 companies analysed was 4% from 2020 to 2024. That is twice the 2% TSR from 2014 to 2018 in BCG’s previous analysis.

But BCG quickly contextualises that “improvement.” Agribusiness still lagged almost every other industrial sector in its comparison set, outperforming only 4 of 36 sectors.

So what happened?

Volatility was real, but the mechanism matters

BCG points to two major drivers of volatility. First, COVID-19 disrupted agriculture supply chains. Next, the war in Ukraine disrupted grain and natural gas supplies, spiking prices for some commodities and inputs, such as fertiliser.

They also add a third structural layer, which is rising variability in weather patterns and climate risks that affect the prices of commodities such as coffee, cocoa, and palm oil.

These shocks helped some companies, especially early in the period. For major row crops, the effects were strongest in the first half, then became more mixed as prices normalised.

Growth happened, but valuations fell

Median revenue grew 8%. Margins and cash contribution rose slightly. But valuation multiples fell by 5% overall, with every subgroup contributing to the decline. A valuation multiple is the price investors are willing to pay for a company’s earnings. When multiples fall, it means the market is paying less for each dollar of profit.

So even though companies grew, investors did not reward that growth with higher valuations. Returns stayed modest.

This is typical in cyclical industries because growth and valuations often move in opposite directions. Investors tend to “price in” expected cyclical movements based on previous cycles.

The implication is important. You can’t rely on growth alone to drive returns. If the market believes your growth is mostly cyclical, it will treat it that way.

What you need is a mix of levers that are credible even when the cycle turns. Margin resilience, cost discipline, balance sheet flexibility, and a clear story about why your earnings power is structurally improving.

Where value was created

BCG splits the sector into five subsectors: agricultural equipment, crop protection and seeds, fertilisers, plant processors, and protein processors. The same shocks played out differently depending on where a company sat in the value chain.

Agricultural equipment led in value creation. BCG notes major players used M&A and joint ventures to upgrade capabilities. Even as unit volumes declined after the 2022 peak, higher prices linked to new technologies supported growth and returns.

In crop protection and seeds, performance diverged. A seed company (KWS) performed well due to margin and revenue growth. Corteva benefited mainly from seeds. Pure-play crop protection companies struggled later in the period as channel inventories were high and raw material supply was abundant, which pushed pricing down.

BCG also notes that large M&A deals in seeds and crop protection have had mixed results, which has pushed some companies to rethink portfolios and move toward simpler, more focused structures.

In fertilisers, BCG notes manufacturers generated the second-highest TSR. BCG cites CF Industries benefiting from lower natural gas prices in the US than in Europe.

In plant processors, results depended on product mix and geography. Ingredient-focused companies did better than commodity processors, partly because exposure to consumer pricing helped offset cyclical swings.

In protein processors, demand grew, but high feed costs limited gains. Higher TSR depended on cost control. BCG also notes growth in seafood processors, with Norwegian salmon companies benefiting from strong demand.

What separated the top value creators

The top ten performers delivered a median TSR of 15%, and every subsector had at least one standout. Growth was the largest driver, but the best performers were able to create value through more than one route, and to some degree offset the usual pattern where strong growth comes with falling valuation multiples.

BCG also points to common traits: market leadership in their subsectors, focus on core business, and aggressive cost and margin management.

Deere stands out. From 2020 to 2024, Deere delivered ~21% per year in shareholder returns. Revenue and margin growth mattered, but Deere was also the only company among the top ten whose valuation multiple increased, and the only company in the full sample that improved across every TSR dimension BCG measured.

BCG attributes part of Deere’s performance to “smart industrial strategy,” and notes it was on track to meet or exceed 2026 targets for connected machines and digital engagement.

The winners “stack” advantages that defy the cycle

McKinsey and BCG are looking at different time horizons, but they are converging on the same meta-lesson. In agrifood, outperformance isn’t about “outguessing the cycle”. It’s about building an operating system that compounds through volatility and earns the right to invest when others pull back.

Here are the three takeaways that stood out to me:

1. Capital efficiency is the real differentiator

In a cyclical sector, capital efficiency is your primary source of optionality. When the cash conversion cycle drifts out by 30 days, you lose flexibility exactly when flexibility is most valuable. The winners treat capital turns as a strategic weapon.

By maintaining control of inventory and unpaid invoices and using assets efficiently, they keep cash available. That cash can fund productivity work, R&D, and selective acquisitions when competitors are playing defence.

2. Growth without structural change is a valuation trap

BCG’s data is a useful reality check: revenue can grow while valuation multiples compress, leaving TSR mediocre. That’s the cyclicality tax. If markets believe growth is just a price wave, they discount it.

To earn a higher valuation, companies need to show that profits are improving for structural reasons. In practice, that means stacking three levers: margin resilience (real differentiation customers will pay for), cost discipline (repeatable productivity), and balance sheet agility (working capital as a core competency).

3. The operating system is a set of repeatable behaviours

McKinsey’s “five big moves” are habits: doing regular, smaller M&A instead of one giant bet, shifting investment away from legacy businesses when returns fade, and investing through downturns rather than waiting for certainty.

Across subsectors, the consistent winners behave like rational capital allocators under stress. They shift from being price-takers in a commodity market to value-makers by continually re-shaping where they play and how they win.

Interested in more Better Bioeconomy content? I’ve been running an agrifood investor interview series, with conversations featuring investors from PeakBridge, Big Idea Ventures, Ajinomoto Group Ventures, and more. More to come in 2026!

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Thanks for being here. For Issue #130 of Better Bioeconomy, I sat down with Oscar Ramos, Managing General Partner at Orbit Ventures, an emerging-markets venture platform that backs breakthrough companies and helps them scale across Asia, Africa, the Middle East, and Latin America.

Orbit doesn’t operate like a typical venture fund. It invests, but it also acts as a hands-on growth platform, helping companies navigate scaling in markets where infrastructure is uneven, buyers are fragmented, and execution matters far more than narrative.

Oscar is open about not being a food or ag specialist investor, which is precisely why I wanted his perspective. It sits outside the usual industry echo chamber. What he brings instead is pattern recognition from working across sectors and geographies where productivity gains, not sustainability slogans, determine whether a business survives.

Before Orbit, he worked across R&D and telecom at companies like Ericsson and Telefónica, then went on to build and launch several startups before moving into venture capital at SOSV, where he remains a Venture Partner. That background places him at the intersection of operators, investors, and founders building under real constraints.

In our chat, we spoke about why logistics and first-mile coordination are often the highest-leverage starting points in agrifood, and why adoption in emerging markets follows immediate economics rather than ideals. We also discussed how the same bottlenecks repeat across continents, what founders can reuse as a result, why localisation determines whether solutions stick, and why the unwritten rule in climate tech in emerging markets is still “don’t talk about climate.”

Let’s jump in!

Productivity is the fastest path to both growth and decarbonization

For Orbit, productivity is a key mandate. Climate sits inside it, not alongside it. When Oscar looks at agriculture and food systems in emerging markets, he is not starting from emissions targets or sustainability frameworks. He starts with a more straightforward question: where is value being lost today, and how quickly can it be recovered?

The thread running through his perspective is inefficiency. The systems around farmers, such as transport, storage, aggregation, and distribution, still run as they did decades ago. Fresh produce often deteriorates before it reaches buyers because basic first-mile transport is inefficient. Weak coordination, empty backhauls, and layers of unnecessary intermediaries add friction at every step, generating more waste and a larger carbon footprint.

The first leg of the journey becomes a kind of operational black hole: loose communication between farmers, traders, and transporters, inconsistent handling, long delays, and avoidable losses. As a result, supply chains burn time, fuel, and product before anyone captures real value.

Oscar sees that waste as the low-hanging fruit. He shared the example of Ridelink, an Ugandan company, backed by Orbit, that coordinates informal trucking networks in rural areas. Instead of building new infrastructure from scratch, the platform makes the existing system legible. It aggregates information on which vehicles have capacity moving in and out of villages, then matches that capacity to farmers and distributors who need it.

The immediate effect is fewer dead-head trips. Drivers earn more because they are not returning empty. Farmers pay less because loads are consolidated. Ag-input distributors reach more customers with the same fleet. The same products move faster and more cheaply, and less fuel is burned along the way.

He sees a similar pattern in seed and input supply. In many markets, farmers still travel long distances to buy inputs from fragmented, informal retailers. If a startup can aggregate demand and coordinate deliveries, those trips collapse. Farmers get hours back, not just money. Distributors expand reach without expanding assets.

What looks mundane on the surface starts to compound. Better routing lowers the cost per trip by cutting unnecessary kilometres, which reduces fuel use and vehicle wear. Vehicles are used more efficiently because they move fully in both directions. Inputs arrive sooner and in better condition, so less is wasted before it is even applied.

Crucially, these gains show up immediately. And that timing is what determines whether a solution is adopted at all.

Drive adoption by showing immediate value

Adoption follows economics. In more mature markets, it is easier to talk about sustainability and long-term gains. In emerging markets, as Oscar put it, “in very traditional links you need to prove your value quickly.” This matters most when solutions are sold to the smallest players in the value chain: smallholder farmers, truck drivers, and rural traders who operate week to week rather than quarter to quarter. A product that promises higher yields in three years will struggle against one that drops operating costs today.

For founders, this means designing offerings around immediate incentives. In the Ridelink example in Uganda, the platform delivered value to all participants quickly: drivers earned more, farmers paid less, and distributors sold more inputs. That quick feedback loop accelerates adoption and builds the trust needed for longer‑term changes, such as adopting improved seeds or on‑farm technologies.

Oscar also emphasised that quick wins don’t preclude long‑term impact. Reducing waste and moving goods more efficiently also reduces emissions. In many supply chains, first-mile leakage is a meaningful emissions driver. In emerging markets, where there is little room for purely altruistic climate initiatives, climate solutions have to hitch themselves to cost savings to gain traction.

He pointed to similar dynamics in Bangladesh. Orbit-backed Fashol has reportedly reduced farm-to-distribution waste from around 50% to under 5% by tightening coordination across the chain. The details vary by market, but the principle is consistent: when a solution quickly turns leakage into cash, adoption follows.

Subscribe now

Most continents share the same bottlenecks, even if the surface looks different

If you are not on the ground, working across Africa, Asia, and Latin America, the differences may seem daunting. Languages, cultures, regulatory frameworks, and infrastructure levels vary widely. Yet Oscar argued that there are more similarities than founders realise.

He recalled two Latin American founders from an Orbit-backed company who had previously helped scale an online marketplace for fresh produce across Brazil and Mexico, and brought those hard-earned lessons into their next venture.

When they compared notes with Orbit’s Nigerian portfolio company, Pricepally, which is working on farm-to-table distribution, the overlap was striking. Pricepally’s team has also learned from Frubana's experience in Latin America (Orbit did not invest in Frubana but has backed former employees building new ventures), which helped sharpen what to replicate and what to avoid. They began to share lessons from their own failed experiments, including which pricing models resonated and what mistakes to avoid when entering new regions.

Similarly, one of Orbit’s Colombian investment team members recently joined the evaluation of a prospective investment in Egypt. The team was surprised by how many of its Colombian playbook elements, such as the need for working capital solutions for informal traders and the importance of transparent pricing in commodity markets, applied directly.

This cross‑market pattern recognition is valuable because it reduces the “reinvention cost.” If you know that first‑mile aggregation is a bottleneck in multiple continents, you can draw on proven solutions rather than starting from scratch.

By cultivating a community where founders share experiences across borders, Oscar believes investors can accelerate learning across markets. For investors, this pattern recognition also reshapes risk. Lessons learned in Nigeria can inform decisions in Bangladesh or Peru, allowing teams to move faster with more confidence.

Localisation is what makes solutions stick

The flip side of these similarities is the risk of copy-and-paste. Not every solution travels seamlessly, and local execution remains crucial. Despite the shared bottlenecks, Oscar warned against assuming that data and logic alone can override centuries of tradition.

Agriculture is deeply cultural. Practices such as harvesting rituals, planting schedules, or animal rearing methods are often tied to identity and pride. In some cases, those traditions persist even when alternative approaches are demonstrably more efficient. This has direct commercial implications. Oscar recounted instances where startups tried to introduce new storage systems or cropping patterns only to face resistance because the methods clashed with cultural norms.

Even when the science proved the innovation would reduce losses or increase yields, adoption lagged until the solution was reframed to respect local identity. Sometimes that meant allowing a traditional practice to coexist alongside a new technique. Sometimes it meant adapting the product so it complemented, rather than replaced, the status quo.

This balancing act echoes a broader tension between efficiency and identity. Many emerging‑market agricultural practices evolved for sound historical reasons: certain crops are planted at specific times to avoid pests, and grains are dried in particular ways to suit local cuisines.

While modern techniques may appear ‘more efficient’ on paper, they must be introduced with humility. In Oscar’s words, “Even if it’s not the most efficient way, it’s part of the identity of the people.” As founders designing technology for agrifood, it is essential to remember that emotional and cultural considerations matter as much as unit economics.

Navigating exits: ‘The unwritten rule of climate tech in emerging markets: don’t talk about climate.’

Finally, we turned to exits. A topic top‑of‑mind for investors as IPO windows remain shuttered and M&A appetite wavers. Oscar was candid: in emerging markets, there are very few clean tech or climate exits. The pool of venture capital is small, and the subset focused on climate is even smaller. That scarcity shapes how companies position themselves. “I have a joke,” he said. “The unwritten rule of climate tech in emerging markets is: we don’t talk about climate.”

This is about pragmatism. In many markets, climate‑aligned products are perceived as expensive, unreliable or time‑consuming relative to conventional alternatives. Meanwhile, investors outside climate circles are sceptical of the revenue potential and timelines of climate businesses.

Founders reframe their solutions as logistics or mobility plays, not as climate. A startup that reduces fertiliser waste by optimising application is positioned as a precision agriculture company that improves productivity rather than as an emissions‑reduction tool. A firm building cold‑chain infrastructure is pitched as a supply‑chain efficiency company. The climate benefits are real, but they are not the headline.

This positioning makes sense given the current market. Global data show that climate‑tech funding has cooled: investment in the sector continues to decline. Exits are largely happening via bargain‑hunting acquisitions rather than blockbuster IPOs.

For companies operating in emerging markets, where the buyer universe is smaller and macro risks loom larger, the more realistic path is to build a real business with strong unit economics (surprise!) and position it as a valuable asset for a strategic buyer, rather than optimising for a public listing.

This pragmatic stance shapes how founders should design their companies from the outset. Businesses that can be read through multiple lenses tend to have more exit optionality. A supply-chain optimisation platform might appeal to logistics operators, commodity traders, or even government agencies. A digital marketplace could be relevant to regional retailers or telecom groups. Keeping the business model flexible widens the pool of potential acquirers.

It also puts pressure on the fundamentals. Strategic buyers will ultimately look past the climate narrative and focus on margins, customer retention, and the ability to scale. If those elements are weak, climate credentials alone are unlikely to carry a deal over the line.

Finally, capital structure matters. As dedicated climate tech capital pools tighten, equity alone is often insufficient, especially for businesses with infrastructure or asset-heavy components. Debt, revenue-sharing arrangements, and blended finance are increasingly part of the picture. Designing a capital stack that matches the realities of the business is becoming just as important as designing the product itself.

Takeaways from my conversation

1. Solve for inefficiency first. Supply chains in emerging markets leak value through poor first‑mile logistics, lack of cold storage and fragmented distribution. Start by coordinating transport, aggregating demand and reducing waste. These are immediate levers for productivity, profitability and emissions reduction.

2. Value drives adoption. Farmers and drivers adopt new tools when they see a quick, tangible benefit. Design solutions that cut costs or raise incomes in the first week, not the first year. Climate benefits are a bonus, not a primary selling point.

3. Look for patterns across markets. Nigeria, Colombia, Vietnam and Egypt share more agricultural bottlenecks than you’d expect. Founders who learn from peers in other regions avoid re‑inventing the wheel. Investors can accelerate innovation by fostering cross‑border knowledge networks.

4. Respect local identity. Efficiency isn’t everything. Traditions and cultural practices shape how agriculture is done and how new technologies are adopted. Innovations must be adapted to local contexts and introduced with empathy.

5. Frame your business beyond climate. In emerging markets, climate funding and exits are scarce. Position your company as solving logistics, distribution or productivity problems. Ensure unit economics are strong so that strategic buyers will pay attention, and be prepared for M&A rather than IPOs.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Thanks for being here. For Issue #129 of Better Bioeconomy, I spoke with Andrew D. Ive, Founder and Managing General Partner of Big Idea Ventures. The fund is one of the most active investors in food tech with more than 150 portfolio companies across 30 countries and a hybrid fund-accelerator model that has shaped much of the category’s early growth.

Before launching Big Idea Ventures in 2018 to “solve the world’s biggest challenges by backing the world’s best entrepreneurs,” Andrew began investing in food startups as Managing Director of SOSV’s Food-X accelerator and has worked across consumer goods, software and entrepreneurship.

In our conversation we walked through why the Big Idea Ventures chose twenty weeks and hundreds of hours with founders instead of the classic twelve week accelerator template, how Andrew is thinking about the next chapter of alt protiens, the three way division of labour between startups, corporates and governments, what a good exit looks like in food and ag, and where the fund is leaning now in its own thesis.

Let’s jump in!

How 800 hours of hands-on work with a startup builds conviction

Big Idea Ventures built its model around a simple idea: you only really understand a team by spending time with them. That is why their accelerator runs for twenty weeks and why they commit hundreds of hours to each company before making larger investment decisions.

Across the program and diligence, the fund spends about ~800 hours with every startup. The first cheque is modest at about 125,000 USD. The real investment is the five months of work the founders and the investment team do together. The larger cheques come later, after the fund has watched the team operate over that period.

This time gives Andrew’s team a view that most early-stage investors do not have. In Andrew’s view, traditional early-stage deals in food and ag often follow a similar pattern. A fund meets a team a few times, sees a deck, asks tough questions in a boardroom, does market and competitor analysis, and then decides whether to write a cheque after “maybe 15 hours of real interaction.”

In those meetings, everyone is on their best behaviour. The founder has rehearsed their answers. The team looks aligned, at least on the surface. You do not see what happens when things go off script.

Over five months, you see much more. You see how founders react when a large food company disagrees with their pricing. You see what they do when a pilot fails. You see whether they listen to feedback and change course, or nod politely and keep going. You see the small frictions between co-founders when they are tired or under pressure.

Spending that much time with teams has also changed how Andrew thinks about failure. When you think about why companies do not make it, the answers are familiar. Competition heats up. The technology does not work as expected. Customers are slower than planned. Those factors are real. But from Andrew’s experience, many of the hardest problems are human.

The companies that shut down most often did so because the founding team could not hold together. Co-founders who started aligned realised they had very different risk appetites or decision styles. Disagreements on strategy hardened into personal conflict. People who looked strong on paper discovered, slowly and then suddenly, that they did not want to work together anymore.

There was a second, related theme. Teams made up entirely of highly technical, highly credentialled scientists struggled more often with customers and commercial traction. The science might be strong and the IP clean, but there was nobody whose job it was to sell. Nobody focused on building partnerships, managing procurement, or translating the tech into something chefs, farmers, retailers or food service buyers could say yes to.

The teams that did well over time looked more rounded. They still had deep technical skill, but they also had at least one person who spent time with customers and partners. Someone who could explain the biology or process in plain language and bring back uncomfortable feedback from the market.

By the end of 20 weeks, the fund can form a deeper view of each company. Some teams look like clear magnets for future capital. They are learning fast, signing the right partners and handling setbacks well. Some teams have potential but need more time and support. Others may become decent businesses, but are unlikely to reshape the food system in a way that fits the fund’s mandate.

The fund is called ‘Big’ Idea Ventures for a reason. It exists to back companies that can make a meaningful dent in how food is produced and consumed, and ultimately some of the most important challenges we’ll experience as a species. The 20-week model and the choice to risk time before major capital are part of how they sort that.

Plant‑based needs better products, while fermentation and cultivated need time and scale

Alt protein is a space Big Idea Ventures knows well. Across its funds, it has backed plant-based, fermentation and cultivated meat companies and has seen the sector move through hype, growth and a more sober phase. The question now is which parts need a real rethink and which simply need more time and scale.

On plant-based, Andrew is clear that the sector is in an in-between place. For several years, companies said they could match animal products in taste, texture, cost, and cooking performance. The story was strong. The branding was sharp. Supermarkets and food service signed up.

For many mainstream meat eaters, the products did not live up to the promise. A few brands got close in certain formats. Many did not. The gap between expectation and reality was large enough that people tried a product once, felt it was worse than advertised, and did not come back. In some cases, they wrote off the whole category after trying a single product.

Fast capital made this worse. Strong early demand and large funding rounds pushed some companies to ship before their products were ready. Attractive packaging and bold claims could not fix an underwhelming eating experience.

All that said, Andrew does not think plant-based is a lost cause. He thinks it needs two things. First, better products. That means a focus on ingredients, formulations and process. It means fewer rushed launches and more careful, iterative work. It also likely means deeper partnerships between ingredient suppliers and large manufacturers, so that the same building blocks can upgrade many products rather than rely on a single hero brand.

Second, time. Consumers who got burned will not keep trying new plant-based options every month. It may take a few years before they are ready to give the category another chance.

He also sees room for a shift in focus. Instead of chasing only one-to-one copies of burgers or nuggets, plant-based companies can build ingredient systems that improve many foods. Fats, proteins, fibres and other components can improve nutrition and sustainability across the board, without always being front and centre on the label.

Fermentation and cultivated meat sit on a different timeline. “The first major question for these two technologies was: Can we do this at all? Can we grow cells or use precision fermentation to produce specific proteins in a controlled manner? The answer today is yes.”

The second question concerns cost. Early cultivated meat work relied on expensive ingredients such as fetal bovine serum. In recent years, companies working on new growth media and process improvements have driven those costs down by large factors. Precision fermentation has also benefited from better strains, smarter process design and cheaper feedstocks.

Now the hard part is scale. The issue is not whether these methods work in principle. It is whether they can operate at volumes and costs that make sense in the real world. Here, regulation matters. Building a major plant can take 100-300 million dollars or more. It is hard to justify that if you can only sell into one small market, or if approvals are uncertain. Countries are moving at different regulatory speeds. Companies cannot sensibly plan mega scale on such a patchwork.

Andrew is still positive about the role fermentation and cultivated systems can play. He believes they can, over time, reach cost levels that compete with some conventional animal products, particularly in specific use cases and formats. He does not expect them to replace industrial animal farming in his lifetime fully. He expects a long period during which both systems will sit side by side.

Subscribe now

When startups, corporates and governments do what they’re good at, innovations scale

Big Idea Ventures sits at the centre of a relationship that spans three very different types of organisation.

On one side are startups. On the other hand, there are large corporates such as Tyson, Bühler, Givaudan and AAK, who anchored Big Idea Ventures’ early funds and still work closely with portfolio companies. A third comprises governments and sovereign investors, who care about food security, jobs, climate targets, and industrial positioning. Andrew’s framing is that each group has strengths and weaknesses that fit together when used in the right way.

Startups are good at moving fast. They explore new science, test new business models and get direct feedback from customers. They accept that many experiments will fail. That is part of their job. They take on scientific and market risk. They are not good at scale. They usually lack the capital, systems and experience to operate many factories or serve every major retailer.

Corporates are strong where startups are weak. They know how to run large plants, maintain quality, distribution and handle big customers. They have teams for sales, logistics and compliance. They know how to operate across countries and channels. They are slower at early-stage innovation. Their structures and incentives make it hard to run many risky experiments. They also have existing products and relationships to protect.

In the best cases, these two sides work together. A startup proves that a product works and that early adopters and new customers want it. A corporate then helps make and distribute it at a scale the startup could not reach alone. If the deal is structured well, both benefit.

Governments and sovereigns add another layer. They do not run startups or corporates, but they set the environment. They decide how much funding to give to early research. They design grants and loan programs for plants and infrastructure. They set rules on safety and labelling that affect whether new products are feasible.

When things go well, each group does what it does best. Startups take early technical and market risk. Corporates pick up what works and push it through global systems. Governments build and maintain an environment where both can succeed.

Big Idea Ventures’ role in this is both investor and translator. Its LP base includes corporates and a sovereign investor, and its team works daily with founders. A large part of the job is helping these groups understand each other and making sure promising technologies do not get stuck between the expectations of a corporate, the reality of a startup and the processes of a government.

What does a ‘good’ startup exit look like?

In food and ag, most exits are still trade sales. The Beyond Meat and Oatly IPOs sit in people’s minds because they were media moments, but that is not the norm.

When Andrew thinks about good exits, he starts with reach and impact. A strong outcome is one where the technology gets into the world at scale, and the acquiring company actually helps that happen. The financial result then follows from that.

He sees two common early exit patterns that are less attractive. One is excitement-driven. A large food company acquires a very young startup because it feels new and interesting. On paper, the deal looks promising. In practice, the product is not ready, the team is not settled, and the startup is absorbed into a large structure that cannot give it what it needs. The technology loses momentum.

The other is distress-driven. This is more common these days. A startup runs out of cash and looks for a home. A buyer steps in, often on terms that are better than a shutdown but not great for early investors. The technology may survive, but usually in a narrow way.

The path Andrew optimises for is more deliberate. A company proves that its technology works. It builds a strong team. It finds repeat customers. It becomes a real business with revenue, processes and a clear role in the market. At that point, if a larger company acquires it and uses its own footprint to take the products into more markets, the result can be powerful for everyone.

This view links to how Andrew thinks investors should behave after they write the first check. Before investing, you care a lot about valuation and terms. You negotiate to a place that is fair for both sides. Once the deal is done, you stop focusing on the paper and focus instead on helping the company succeed. You expect to be in the story for six to ten years. You accept that you do not control the exit. You support the founders in building a business that many potential buyers would want to buy.

The advantage a good investor has, in Andrew’s view, is not superior intelligence. It is pattern recognition and distance. Founders live inside the company every day. Investors see across many companies and can help with context, introductions and framing. If an investor behaves as if they are smarter than the founder and tries to run the company by remote control, they will make things worse.

Zooming in on the building blocks of food innovation

Where does Andrew see opportunity today? One clear focus is ingredients. After its early work in alt protein, Big Idea Ventures has spent more time with companies working on fats, flavours, colourants, sweeteners and innovations around crops like cocoa, coffee, sugar and palm. The appeal is that ingredients are modular. A better fat system can improve many products. A new flavour approach can shift taste and nutrition at the same time without relying on one flagship brand.

Another area is food waste upcycling, but with more technical depth than before. Early upcycling often meant repurposing, such as turning spent grain into flour or snacks. Newer companies are using biotechnology to turn specific waste streams into higher-value ingredients. They work on defined outputs and often aim for clear functional roles in food formulations.

Fermentation remains central, with growing interest in bacteria-based platforms. Bacteria can grow faster than yeast in many cases. They can sometimes use a wider range of low-cost feedstocks, including byproducts and wastes. That combination can help both economics and climate performance, especially when paired with upcycling themes.

Closely linked to this are tools that search for better biology. Instead of improving only one strain at a time, some companies are running structured competitions among multiple strains of algae or microbes. They use data from those competitions to select the best performers. This moves the work from one-off experiments to a more systematic search.

The fund is also increasing its efforts to partner with university research. Generation Food Rural Partners (GFRP), created and managed by Big Idea Ventures, licenses and commercialises university intellectual property in food tech, agritech and sustainable materials from a growing network of more than 30 partner universities across the United States and with plans to expand this model into Europe in 2026.

There is also a clear geographic angle. Andrew and his team are paying closer attention to Japan and other parts of Asia, as well as to some markets in the Middle East. In these places, food security, climate pressure and economic diversification are all strong drivers. As with Singapore more than five years ago, Big Idea Ventures likes to be early to new innovation hubs around the world, helping to build a healthy, thriving food and agricultural innovation ecosystem. That often means more policy support and corporate interest for the kind of companies Big Idea Ventures backs.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Thanks for being here. For Issue #128 of Better Bioeconomy, I’m trying something new. I want to experiment with a monthly edition on AI (artificial intelligence) in agrifood, a format that lets me look across the developments from last month to see where patterns are starting to emerge.

Why this format? Because I increasingly believe that intelligent systems, whether we label them AI, ML (Machine Learning) or automation, can meaningfully accelerate progress in agrifood. The exact terminology is less important for this series. I’m using “AI” here as a practical shorthand for that whole toolkit.

So in this issue, you’ll find a curated round-up of November’s AI x agrifood news, a World Bank/Gates report on AI as agricultural infrastructure, and some of my reflections on where the leverage (and the risk) might be.

This is still an experiment, so I’d love to hear what you think. As always, my understanding keeps evolving as I dig deeper. My goal with this series isn’t to declare an AI revolution or to pretend I’m an AI expert. If you spot something I’ve missed or got wrong, let me know.

Let’s jump in!

AI X AGRIFOOD NEWS

🐛 Bindwell raised $6M to reinvent pesticides using AI

• Bindwell is reframing pesticide discovery as an AI-native search problem, adapting drug-discovery models (Foldwell, PLAPT, APPT) to predict protein structures, protein–protein interactions and binding affinities for agricultural targets, enabling rapid in silico exploration of novel modes of action.

• The company reports that its structure-prediction system runs ~4x faster than AlphaFold 3, while APPT achieves ~1.7x higher accuracy on a standard affinity benchmark. Taken together, the “scan every known synthesised compound in ~6 hours,” generating a computational funnel that narrows vast libraries to a short list of viable candidates.

• Its $6M seed round will fund the shift from benchmarking to wet-lab validation, with Bindwell intending to license IP for new pesticide actives, positioning it as a next-generation agrochemical discovery engine.

Source: TechCrunch

🍫 World’s largest B2B chocolate maker is partnering with NotCo to bring AI into chocolate formulation

• Barry Callebaut is integrating NotCo’s Giuseppe AI into its chocolate R&D pipeline, combining the startup’s ingredient-chemistry and sensory-modelling capabilities with the manufacturer’s extensive formulation data to modernise innovation amid cocoa volatility, climate stress and health-driven reformulation pressures.

• The AI analyses product recipes and ingredient data to predict flavour, texture, and functional outcomes, and proposes recipes that can reduce cocoa usage, improve nutritional profiles, or meet Nutri-Score frameworks.

• By shifting early-stage formulation into computation, the partnership aims to cut trial-and-error cycles, increase hit rates and accelerate product development, supporting a new generation of chocolate that maintains indulgence even as formulations evolve for cost, resilience and health considerations.

Source: Food Ingredients First

♻️ MOA Foodtech’s AI fermentation service turns food waste into sustainable ingredients

• MOA Foodtech’s “MOA Box” applies AI-driven bioprocess design to help manufacturers convert starch-rich byproducts into high-value protein ingredients through fermentation, enabling clients to access advanced bioprocess R&D without building internal infrastructure.

• Its Albatros platform runs 300+ bioprocess simulations per hour, matching each byproduct stream with the optimal yeast strain and fermentation parameters to identify the most efficient, low-cost, and low-impact route. MOA then executes fermentation, downstream processing, and validation before licensing the turnkey process back to the client.

• The company reports biomass with all essential amino acids and a protein digestibility score of 0.9 (on par with soy, egg and beef). It claims the approach can generate up to 17.5x more value from starch byproducts while taking clients from byproduct assessment to a market-ready ingredient in ~6 months.

Source: Green Queen

🌾 Avalo harnesses ‘whole genome AI’ to accelerate crop breeding

• Avalo applies “whole genome AI” to crop breeding by analysing entire genomes alongside environmental and field-performance data, allowing its interpretable models to identify multi-gene patterns linked to complex traits like drought tolerance, nutrient efficiency and yield stability.

• This AI layer sits within a rapid-evolution platform that creates diverse genetic populations, models their performance computationally and then directs breeders toward the most promising crosses, effectively simulating much of the breeding pipeline in silico before committing to long, resource-intensive field trials.

• Avalo says this system can halve breeding timelines for crops as complex as sugarcane and highlights early proof points such as a cotton line requiring zero irrigation and 70% less fertiliser, suggesting the platform can generate more climate-resilient, resource-efficient varieties than conventional breeding pathways.

Source: AgFunder

🐕 Cargill put an AI-powered ‘robot dog’ on the factory floor

• Cargill has deployed a Boston Dynamics Spot robot at its Amsterdam oilseed crush plant, where it performs roughly 10,000 autonomous inspections per week, using computer vision, thermal imaging and onboard ML to navigate the facility, detect overheating, leaks and other anomalies, and feed structured insights into plant monitoring systems.

• The company says Spot has already identified issues such as an overheated decanter and bearings oscillating between 40°C and 100°C, enabling interventions before equipment failure and unplanned downtime. It reports clear gains in early detection, reliability and safety, with the robot delivering a level of consistency that a human walkaround cannot match.

• Rather than treating Spot as a labour substitute, Cargill is retraining staff to design missions and interpret the robot’s data, framing the shift as a move toward predictive, data-informed operations. The pilot remains limited to one European site, but the firm is assessing a broader rollout across its network.

Source: AgFunder

😋 Food System Innovations receives $2M to build an open-source AI model for flavour and texture in alt-proteins

• Food System Innovations’ NECTAR program has received a $2M Bezos Earth Fund grant to build an open-source AI model that helps companies design better-tasting sustainable proteins, drawing on work led by Stanford computer science PhD researcher Anna Thomas.

• NECTAR has amassed a large dataset pairing consumer sensory responses to alternative proteins with detailed molecular and ingredient data. The AI is being trained to connect molecular structure, flavour, texture, and preference patterns so it can predict which formulations are most likely to succeed before companies commit to costly development cycles.

• The project’s stated benefits include enabling CPGs and manufacturers to iterate faster and more cost-effectively in alt protein R&D, improving the likelihood that new products succeed on taste, and supporting emissions reduction by making plant-based and other climate-friendly proteins more competitive with conventional meat “one bite at a time.”

Source: Green Queen

📚 AKA Foods raised $17.2M to launch an AI platform that restructures fragmented R&D into an actionable knowledge base

• AKA Studio is a domain-specific AI platform that ingests a food company’s decades of scattered R&D, sensory and regulatory data and restructures it into a unified, searchable “food grammar.” It captures ingredient specs, experimental measurements and sensory attributes like texture, aroma and taste in one coherent knowledge base.

• AI assistants and agents sit on top of this foundation, querying internal and external data to recommend formulation improvements or new product concepts, giving R&D teams an intelligent co-pilot that can reason across chemistry, functionality, sensory outcomes and regulatory constraints without acting as an opaque black box.

• AKA Foods says this approach can shrink innovation cycles from years to weeks, raise the quality and confidence of R&D decisions, and support the development of cleaner-label, lower-sugar, lower-fat and more resilient formulations, all while ensuring strict data security through private or on-prem deployments where client data never leaves their environment or trains shared models.

Source: FoodBev

🤖 Ceres AI raised $13M to scale an agricultural intelligence platform spanning 32M acres

• Ceres AI has raised $13M to scale its “AI for Agricultural Intelligence” platform, which is built on 17 billion plant-level measurements spanning 32 million acres and 40+ crop types. Its models analyse this dataset, turning imagery and plant-level measurements into high-resolution analytics for agribusiness and financial institutions, supporting insurance, lending and other risk-management workflows.

• These insights feed into the workflows of agribusinesses, insurers, lenders and farmland investors, where the AI now informs underwriting, loan pricing, risk monitoring, input allocation and portfolio management. The company positions this as an emerging “agricultural intelligence layer” that makes decisions faster, more consistent and more transparent.

• Ceres AI claims this infrastructure can improve performance, profitability and climate-risk management for customers. The round also drew attention for the appointment of Arista, an AI agent, as what they describe as the first AI board member.

Source: iGrowNews

Subscribe now

AI X AGRIFOOD REPORT

Harnessing AI for agricultural transformation

• A World Bank-led report argues that AI is becoming a foundational tool for agricultural transformation, especially in LMICs (Low- and Middle-Income Countries), where climate shocks, rising input costs and weak supply chains threaten food security. It highlights that AI’s value lies in improving productivity, sustainability and inclusion across agrifood systems.

• AI is positioned as uniquely suited to agriculture because the sector generates multi-modal, unstructured data, from images and text to climate records and sensor inputs. Generative AI builds on this by enabling localised, conversational advisories that can overcome literacy, language and access barriers for small-scale producers.

• The report also highlights high-impact use cases across the value chain. This includes crop and livestock discovery, AI-powered advisory and farm management, inclusive finance and risk mitigation, market and logistics optimisation, and cross-cutting applications such as satellite-enabled crop forecasting and agro-ecological zoning.

• It stresses that AI will only deliver equitable benefits if supported by foundational enablers: robust connectivity and energy infrastructure, localised datasets, digital literacy and human capital, clear data governance, and strong public-private ecosystems. This includes building agricultural Digital Public Infrastructure (DPIs) like farmer registries, data exchanges and open service networks.

• The investment priorities of the report focus on agriculture-specific AI models, foundational data systems, compute infrastructure, and policy frameworks that enable responsible scaling. The forward look warns that AI must be guided by inclusion, ethics, and local relevance to avoid reinforcing inequalities while unlocking climate resilience and productivity gains.

Source: World Bank

REFLECTIONS

Emerging patterns

1. AI is becoming a workhorse in discovery

Bindwell, Avalo, MOA Foodtech, the NECTAR project and the Barry Callebaut-NotCo collaboration all point in the same direction: AI is being treated as a viable tool inside discovery workflows. The common thread is that these teams are tackling problems where the search space is too large for human iteration alone – chemical space for pesticides, combinatorial breeding space for complex crops, the huge design space of fermentation conditions or flavour systems.

If platforms like Bindwell move beyond benchmark performance into validated hits, and if Avalo’s cotton result proves reproducible, the front end of the R&D cycle could compress significantly. That still leaves regulatory and manufacturing timelines intact, but it meaningfully shifts which projects become economically viable to pursue.

A notable feature of this wave is how domain-specific the systems are. Bindwell’s models are tuned for agrochemical binding, not generic protein folding. Avalo’s models are wired into specific crops and environments. NECTAR is explicitly built on sensory panels and meat-eater preferences, while NotCo’s Giuseppe is trained on food composition and processing constraints. The leverage comes from combining solid domain data, clear objective functions, and an architecture that respects the constraints of regulation, scale-up and cost.

2. Food companies are turning decades of R&D into an asset

AKA Foods, Symrise and, indirectly, the Barry Callebaut–NotCo deal all highlight another shift: large food companies are starting to treat their messy archives of R&D reports, sensory tests and regulatory dossiers as something that can be structured and queried. AKA’s “food grammar” framing captures this well: the goal is to encode the relationships between ingredients, processes, sensory outcomes and constraints so an AI system can reason with R&D teams rather than hallucinate novelty.

This is less glamorous than “AI chef”, but it lines up with how big organisations typically work. Formulation teams are constrained by past experiments, existing supply contracts, compliance rules and brand guardrails. Systems like AKA Studio are essentially knowledge-operating systems for that reality.

If they work, the payoff is not that AI suddenly “invents” a new category (although that might be possible), but that teams can avoid repeating old mistakes, respond faster to shocks (like cocoa price spikes) and have a clearer map of what is technically possible within their own house.

3. Risk, finance and infrastructure are becoming natural homes for AI

Ceres AI sits firmly inside the story the World Bank report is telling: AI as an “invisible” layer that informs credit, insurance and policy decisions rather than a frontline app. There’s a clear logic to concentrating AI effort where data is already aggregated – across millions of acres, loan books, insurance portfolios – and where small improvements in risk assessment or pricing can move real capital.

The idea of an “agricultural intelligence layer” isn’t entirely new, but the scale of Ceres’ dataset and the willingness of financial institutions to integrate those signals into underwriting and portfolio management are meaningful signals.

The report’s emphasis on Digital Public Infrastructure (farmer registries, data exchanges, payment rails) also underlines how much of AI’s impact in agrifood is going to depend on unglamorous plumbing. Without reasonably clean, linkable data on who farms where, what they grow and which services reach them, even the smartest risk models or advisory bots are constrained. Ceres, Farmdar-type platforms, and public efforts like KIAMIS in Kenya are all, in different ways, pushing toward that “shared infrastructure” picture.

But this also surfaces a real tension. Systems like AKA Studio, Ceres, and emerging DPI initiatives rely on structured, aggregated data, yet their incentives around openness differ. There is a tradeoff between building defensible proprietary datasets (which investors like) and enabling the interoperability that development institutions argue is essential for equitable outcomes.

Investability

Looking at this through an agrifood tech VC lens, a few areas stand out as more attractive (on paper) for me.

First is vertical, data-rich AI platforms with clear buyers and existing data flows e.g. agricultural risk/finance (Ceres) or enterprise R&D co-pilots (AKA, Symrise-like stacks). They benefit from high switching costs once embedded, measurable impact (loss ratios, time-to-formulation), and relatively clear paths to revenue. The risks are concentration (a few big customers), regulatory scrutiny (especially in finance), and dependence on third-party infra (cloud, DPIs).

AI-augmented discovery for inputs and traits (Bindwell, Avalo, MOA) is also interesting. The upside is substantial if they can own IP around new actives or germplasm. The downside is classic deeptech: high technical risk, long validation cycles, regulatory exposure, and capital intensity in the wet-lab and scale-up stages. These look suitable for patient capital with a strong view on regulatory timelines and exit paths (licensing vs own commercialisation).

On a related note, check out insights from my recent chat with Yoni Glickman, Managing Partner at PeakBridge, who shared how AI actually becomes a value-add in food.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Thanks for being here. For Issue #127 of Better Bioeconomy, I sat down with Yoni Glickman, Managing Partner at PeakBridge, to talk about how agrifood-tech has evolved over the past few years and where the most compelling opportunities are now emerging.

Yoni is one of the rare investors who has built billion‑dollar ingredient businesses and venture portfolios from the inside out. He served as President of Frutarom’s Natural Solutions division when IFF (International Flavors & Fragrances Inc.) acquired the company for $7.1B, one of the largest deals in the history of the flavours and ingredients sector.

He also founded the FoodNext innovation lab and ran companies in water technology and sustainable materials. At PeakBridge, he now sits across both FoodSparks, the seed fund first launched in collaboration with EIT Food, and Growth Fund II, giving him a 360° view of the food landscape.

In our chat, we break down why certain agrifood bets fell short, where the real opportunities are emerging, and where AI actually creates moats. We also dig into how PeakBridge evaluates companies from seed to growth, why unit economics and total capital required matter most, and why strategic M&A remains the most likely exit path.

Let’s jump in!

When great technologies collide with hard economics

If you rewind the last five years of agrifood tech, the pattern is familiar. A big story captures attention, capital floods in, a few companies build real businesses, and most struggle to translate early promise into commercial traction. Yoni puts three themes in the overhyped bucket: vertical farming, insect protein and precision fermentation targeting commodity applications.

Vertical farming, as a category, promised local, pesticide-free greens, shorter supply chains and climate resilience. In reality, the CAPEX and energy burden of running fully controlled environments made it very hard to reach the cost and return thresholds needed for mainstream, commodity-adjacent produce.

A few narrow niches around high-value ingredients can still make sense, since controlled-environment production can work when the economics support it. Still, as a broad thesis, he thinks it is “objectively clear” that it did not work.

Insect protein followed a similar arc. Significant capital went into the sector, but genuine demand never caught up with the capacity being built. Yoni summarises that there is “no real product-market fit,” and the category has often looked like a solution looking for a problem rather than an answer to a clear customer need.

Precision fermentation, in his view, belongs in the “tool, not thesis” bucket. Fermentation has been an “incredibly interesting and relevant tool” for food and ingredients for decades. The misallocation came from using precision fermentation on bulk proteins and expecting them to compete on cost with deeply optimised dairy and commodity plant supply chains. Used for differentiated, higher-value ingredients, the tool still makes sense. But used for commodity proteins, the economics are unforgiving.

Even in AI, which he believes will quietly infuse most of the food system, Yoni is wary of discovery platforms that brand themselves as the “AI engine” for new molecules but rely on long timelines and consulting-style revenue. Discovery work can be valuable, but the combination of long development cycles and business models that do not capture enough downstream value “really don’t stack up as a venture investment.”

The quiet compounding opportunity in food waste and speciality nutrition

While capital chased the stories that made headlines, Yoni sees large, underinvested opportunities in areas that look less glamorous at first glance.

Food waste is one of them. It still shows up in many sustainability decks as an ESG talking point, yet structurally it sits at the intersection of food security, inflation and margin. A third of all food produced globally is never eaten, and a meaningful slice of that loss occurs in hospitality and foodservice, where every kilo wasted is wasted CAPEX, labour, energy, and ingredients. Treating food waste primarily as a reputational issue misses how much it matters to the P&L.

Yoni referred to Orbisk, one of PeakBridge’s portfolio companies. The company outfits professional kitchens with a fully automated food waste monitor: a smart camera and scale installed over the bin that uses AI to recognise what is being thrown away and in what quantity, then streams that data into a dashboard with clear recommendations.

Kitchens do not have to change their workflow. They simply throw food away as usual and wake up to ingredient-level insight into what they overproduce and when. For customers, the result is less waste, better purchasing and menu decisions, and savings that comfortably exceed the subscription fee.

The second gap sits in specialty ingredients and what Yoni frames as the “boring” parts of the stack. Rather than building brand-new commodity categories, he is drawn to scientifically substantiated nutrition ingredients, flavour systems, and other functional components used at low inclusion rates but with an outsized impact on performance, stability, taste, or health.

Many of these businesses rely on traditional manufacturing methods such as botanical extraction or solid-state fermentation. They are familiar to industry buyers and can produce highly differentiated ingredients without requiring exotic bioprocessing infrastructure.

How AI actually becomes a value-add in food

Today, almost every pitch deck claims some AI angle, whether in discovery, supply chain visibility, marketing or operations. For Yoni, AI in food only becomes investable when two conditions are met at the same time: the tool fits seamlessly into existing workflows, and shows measurable ROI for the customer.

Integration comes first. Food and beverage companies operate in heavily regulated environments with established hazard systems, quality routines and compliance processes. Tools that ask teams to completely re-architect how they work rarely get past the first pilot.

AI earns the right to spread when it is embedded inside how procurement, R&D, marketing or operations already function, rather than bolted on as a parallel workflow. Orbisk, as mentioned earlier, is one example of that in foodservice, where teams that throw food away as usual get waste analytics without changing their routine.

The second condition is measurable ROI. AI tools that cannot be tied to clear metrics will eventually be crowded out by projects that can. On the waste side, that might be kilos of food saved per site per year, along with the associated cost reduction. On the commercial side, it might be higher new product success rates, better campaign performance or a shorter concept-to-launch cycle.

PeakBridge’s portfolio company, Tastewise, positions itself as a generative AI (GenAI) platform that converts large, messy streams of food and beverage consumption data into targeted recommendations for product innovation, sales and marketing.

The moat is not that it uses GenAI. It is that it lets global brands like Kraft Heinz, PepsiCo and Mars test concepts against real behaviour data, craft messaging that aligns with how people actually talk about food, and do this at a speed and scale that is hard to match.

Seed vs growth: funds evolve, so does the definition of ‘investment ready’

PeakBridge runs a seed vehicle and a growth-stage vehicle across very different parts of the food system. FoodSparks is the pan-European agrifood tech seed fund managed by PeakBridge and first launched in collaboration with EIT Food. Growth Fund II is their early-growth fund, investing globally from Series A onwards. Together, these two vehicles give Yoni a clear view of how the definition of “ready” changes both across stages and over the life of a single fund.

One point he returns to is that any fund, seed or growth, progresses over time. Startups often do not realise that what an investor can do in year one of the fund is not what they can do in year four. Early in FoodSparks, with more years left until the end of the fund, they could back a strong team, a forward-looking solution for the food industry and a solid proof of concept (POC), even if product–market fit was not yet fully proven.

As the fund matures and the exit window approaches, the same seed fund has to become more conservative. In later years, FoodSparks prioritised companies that had moved beyond POC to real product–market fit and revenue, with a shorter, clearer path to follow-on rounds and potential exits.

At the growth stage, the bar rises again, and the criteria diverge by vertical. For a nutrition or health ingredient, a company that might once have been fundable at seed based on strong preclinical work in vivo or in vitro now needs full human clinical trials, double-blind, placebo-controlled, and scientifically robust.

Similarly, for cell-based or other industrial production technologies, early theoretical work in a bench scale might have been acceptable at the seed stage. But at growth, PeakBridge wants to see full-scale manufacturing with positive unit economics, not just models built on lab-scale yields. Growth vehicles have to be more de-risked by design.

Across both funds, regardless of stage, unit economics and capital efficiency are non-negotiables. First, there must be a realistic path to positive unit economics once you model scale, raw material costs, location and the actual manufacturing approach. Second, the total capital required to reach cash flow positive has to align with venture funding cycles.

Subscribe now

How “food as medicine” becomes proven nutrition, not wellness noise

Food as medicine is having a mainstream moment, helped by the visibility of GLP-1 drugs and a renewed focus on metabolic health. PeakBridge’s interest in this space predates the current wave. Yoni has spent much of his career in “taste and health”, leading natural ingredients and health platforms at Frutarom and IFF and backing botanical and clinically validated ingredients long before public markets were comfortable with that framing.

What separates a functional ingredient that peaks as a wellness fad from one that becomes true B2B infrastructure? Yoni points to three things. First, it must address a meaningful health condition rather than a marginal benefit. Second, it must be backed by strong science, ideally including human clinical trials. Third, it must sit within a supply chain the team deeply understands, from raw materials through extraction, standardisation, formulation, and quality.

PeakBridge’s KÄÄPÄ Biotech is a good example of the kind of business that fits this pattern. KÄÄPÄ researches, cultivates and processes functional mushrooms in vertically integrated facilities, using its proprietary NordRelease process to produce standardised extracts for wellness and nutraceutical brands, supported by clinical data.

EFSA’s formal validation of NordRelease as a non-novel extraction process has helped cement its position as a leading European supplier of high-quality functional mushroom ingredients at a time when the EU market is wrestling with inconsistent imports and variable quality.

Mushrooms may be buzzy, but KÄÄPÄ’s strategy is not built on buzz. It is built on control from spore to extract, rigorous standards for bioactives and safety, and a pipeline of studies that can support health claims in a stricter regulatory environment. That is the profile PeakBridge is looking for in food-as-medicine (or ‘proven nutrition’ as Yoni prefers) more broadly.

Looking ahead, Yoni expects three areas to anchor PeakBridge’s next chapter: proven nutrition, specialty ingredients with clear use cases, and data solutions that integrate easily into food-industry workflows and deliver obvious ROI.

Regulation and CAPEX: design for the world you operate in

If there is one area where Yoni sees otherwise strong founding teams repeatedly stumble, it is regulation. The temptation is to treat it as a late-stage hurdle that can be outsourced to consultants once the product is ready. In reality, regulatory pathways for food, nutrition, and biologically derived ingredients are becoming stricter and vary widely across markets.

Yoni is still surprised by how often founders assume that “natural” automatically means compliant. From his perspective, if a team cannot explain its regulatory pathway and show that the business model remains viable under realistic timelines and scenarios, there is little point continuing the discussion on CAPEX.

Once the regulatory path is clear, the CAPEX stack comes into focus. Here, the differences between categories are large. Consumer brands in beverages or snacks can often rely heavily on co-manufacturing and focus their capital on brand, distribution and working capital.

Ingredients and industrial processes are a different story. Precision fermentation plants are expensive, and while CMOs can work for high-value, low-inclusion ingredients, it becomes much harder to justify for commodity-like products.

By contrast, some ingredient technologies can scale with inherently lower capital requirements. Solid-state fermentation and certain botanical extraction models can leverage established contract manufacturers and equipment suppliers, allowing teams to scale without committing to large facilities.

Yoni notes that several clinically validated ingredient companies have been built entirely on outsourced extraction. This is part of why PeakBridge is comfortable with specific ingredient and functional mushroom plays: not because they are simple, but because their manufacturing pathways are clearer and less capital-intensive.

Building with an M&A endgame in mind

At some point, the question is how all of this value gets realised. Inside PeakBridge, there is a strong internal consensus that most successful outcomes in agrifood will be strategic M&A rather than IPOs.

PeakBridge maps potential acquirers from the investment stage. In ingredients, that might mean global players in flavours, fragrances, nutraceuticals or speciality nutrition. In dairy or protein, it could be the major ingredient companies. On the data and AI side, it might be large tech or enterprise software firms, supply chain platforms or major CPGs looking to integrate an additional tech stack into what they already have.

Different buyers optimise for different things: some are essentially buying market share and EBITDA, others are looking for a strategic technology wedge or differentiated data asset.

The best founders Yoni works with hold two pictures at once. One is their own long-term vision for what the company could become if it stayed independent. The other is a realistic view of who might buy them, why that buyer would care and what capabilities or metrics would make them compelling.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Thanks for being here. For Issue #126 of Better Bioeconomy, I sat down with Hiroyuki Saito, Manager and Asian Region Lead of Ajinomoto Group’s corporate venture capital (CVC) team, Ajinomoto Group Ventures.

Ajinomoto Group is a ¥1.5 trillion-revenue multinational that launched the world’s first commercial umami seasoning in 1909. It has since grown into a global food and biotechnology enterprise spanning seasonings, frozen foods, sweeteners, amino acids, and even electronic materials.

Today, it operates across 31 countries and regions, supplying everything from tabletop seasonings in Asian kitchens to functional ingredients embedded in soups, snacks, ready meals, and beverages made by other brands.

Ajinomoto Group Ventures sits on top of that legacy. The CVC arm invests across healthcare, ICT, food & wellness, and green domains, backing startups that can leverage Ajinomoto Group resources to make an impact.

In our conversation, Saito walked me through how the team defines true strategic fit, the three pillars guiding its agri-food investing, how they balance strategic and financial returns, where they see the next wave of food-health innovation, and what founders need in place before reaching out.

Special thanks to Yuichiro Ohno for helping facilitate the conversation and coordinating internally throughout the process. 🙏🏾

Let’s jump in!

Ajinomoto Group maps its agri‑food bets across three domains

Ajinomoto Group Ventures is a strategy-first investor. For Saito and his team, the startups worth engaging are those that map directly onto the group’s mid-term growth domains. These domains are clearly scoped business areas with internal owners, budgets, and operating mandates.

Within food and agriculture, Ajinomoto Group Ventures is paying attention to three areas:

Green Food

This is where the group focuses its efforts on alternative proteins and next-generation ingredients, leveraging its legacy in taste, texture, and fermentation.

In this pillar, the company has invested in Australia’s v2food to pair its plant-based protein platform with Ajinomoto Group’s “Deliciousness Technology” and global network to make cleaner-label protein ingredients more scalable across markets.

The group is also working with Finland’s Solar Foods to bring Solein, a microbial protein made from CO₂ and renewable energy, into consumer products. Their collaborations sit under Ajinomoto Group’s Atlr.72 brand in Singapore and span early launches across desserts and beverages.

On the coffee side, Ajinomoto Group is co-developing bean-free coffee with Singapore’s Prefer. The partnership allows Prefer’s upcycled, fermentation-derived coffee to be blended into Atlr.72 brand’s GRe:en Drop Coffee line, reducing the drink’s carbon footprint while keeping a familiar flavour profile for local consumers.

Agro-solution

This is Ajinomoto Group’s umbrella for agri-biologicals. The group is building on its existing biostimulant fertiliser products and expanding into a broader set of biological tools.

In Saito’s framing, this spans bio-pesticides, nitrogen fixation, carbon sequestration, and soil analysis and improvement. These are areas where Ajinomoto Group sees biology complementing or replacing conventional agricultural inputs, and where its long history with biostimulants gives it a starting point for deeper collaboration with startups.

Animal nutrition

This pillar is built around Ajinomoto Group’s leadership in lysine formulations for livestock. Its flagship product, AjiPro®-L, is a rumen-protected lysine that improves amino acid balance and reduces greenhouse gas emissions from cattle.

Ajinomoto Group holds a top global market share in lysine additives and is now converting AjiPro®-L’s GHG reductions into carbon credits through Japan’s J-Credit Scheme, targeting one million tons of annual reductions by 2030.

Saito’s team is looking for startups that can complement this strategy, especially those advancing feed efficiency or livestock emissions reduction.

Strategic fit drives investment, but financial returns keep the collaboration alive

Every cheque from Ajinomoto Group Ventures is anchored in the group’s roadmap for where it wants to play and what capabilities it needs to build or access. Within that frame, Saito thinks in two time horizons.

Some investments are made with relatively short-term collaboration in mind, where joint projects and commercial returns can be sketched out early. Others are made for “market sensing,” where the goal is to stay close to an important emerging area with a mid to long-term perspective, even if the collaboration pathway is not yet obvious.

Strategic alignment on its own is not enough, though. Collaboration only compounds if the startup grows, so the team also looks at financial returns and scalability as part of the review. Growth potential and sound economics are treated as prerequisites for any strategic partnership to work over time, not as a separate, purely financial exercise.

Ajinomoto Group’s position in Asia adds another layer. The group enjoys strong brand recognition and deep distribution in the region, which means “the necessity of investment is always a point of discussion internally.” In some cases, Saito says, Ajinomoto Group Ventures may decide not to invest but still move forward through collaboration or a venture-client model, using its channels and capabilities without taking equity.

Because of this mix of tools, he encourages founders to treat the CVC as the front door into the group. The team acts as a primary point of contact, helping startups navigate to the right internal counterparts and shape collaboration structures that make sense for both sides.

In practice, strategic fit, growth potential, and choice of collaboration model are all part of the same conversation. Ajinomoto Group Ventures is investing where it wants to learn and build, but remains selective about when equity is the right instrument and expects any partner to show a credible path to scaling alongside the group.

Ajinomoto Group sees the future of food‑health in tailored diets

When the conversation turned to food and health, Saito pointed to a recent program run by Ajinomoto Group Thailand, the group’s ASEAN hub. That accelerator focused on three themes: personalised nutrition, alternative proteins, and fermentation technology, which now anchor Ajinomoto Group’s thinking at the food-health interface.

The personalised nutrition piece builds on work already in market. In Japan, Ajinomoto Group operates a meal-planning service powered by a proprietary nutrient profiling system, allowing it to generate tailored diets. As D2C and e-commerce channels expand across Asia, Saito sees greater potential to deliver and iterate on these personalised offerings beyond traditional retail.

Alternative proteins and fermentation technologies sit closer to the Green Food pillar, but the emphasis is on how these tools integrate into everyday diets, not on standalone novelty. Ajinomoto Group’s strengths in taste, fermentation, and formulation can help these new ingredients fit more naturally into local cuisines and health needs.

Across these themes, the company is looking for places where its existing capabilities in ingredients, flavour design, and nutrition science can support more tailored, health-aware food services.

Subscribe now

Supporting startups through deep R&D, distribution, and global reach

Ajinomoto Group Ventures’ value to startups goes far beyond capital. Saito describes a system of scientific depth, operational experience, and global infrastructure that few food companies can match.

Upstream, Ajinomoto Group works directly with startups on R&D, bringing together its proprietary bio and fermentation technologies with decades of experience designing “deliciousness” across taste and texture. In areas like cultivated meat, that means supporting media formulation and technical development while also helping make end products more appealing and consistent. The goal is hands-on collaboration that improves both performance and sensory quality.

Downstream, Ajinomoto Group can open sales channels across Asia. Its long history of selling seasonings, ingredients, and finished products into retail and foodservice gives startups a practical route for pilots and early roll-outs. For many companies, testing and scaling with a partner who already understands how to navigate Asian markets is more valuable than a cheque.

For startups working with Ajinomoto Group, that combination of global reach and speed stands out. As Jake Berber, co-founder of beanless coffee startup Prefer, told me, “It’s rare to find a company that is globally impactful, innovative, and works with velocity. They’ve been the perfect partner for us to scale with.”

He added, “Launching our first product with Ajinomoto Group’s Green Business Development Department in Japan and partnering with Ajinomoto Thailand has been an incredible experience, and we’re excited for even bigger collaborations ahead.”

Ajinomoto Group Ventures also supports global expansion. The team has bases in Silicon Valley and Tokyo, and the wider group operates innovation hubs across the United States, Japan, Europe, and other regions, with a presence in 31 countries and territories. These hubs act as sensing and execution nodes, helping portfolio companies explore new markets over time.

Saito’s long-term aim is to connect these hubs more organically so that a startup working with Ajinomoto Group in one geography can more easily extend into others without starting from scratch.

Localisation of food is one of Ajinomoto Group’s greatest strengths

All of Ajinomoto Group’s global machinery sits on top of something older and more fundamental: a hard-won ability to localise food country by country without losing what makes its products work. As Saito puts it, localisation of food is one of Ajinomoto Group’s greatest strengths.

Sensory science provides the universal anchor. Umami is recognised globally as a basic taste, and glutamate activates the same receptors whether you are eating in Japan, Thailand, Brazil, or Nigeria. Ajinomoto Group’s core amino acid and taste technologies can travel.

What changes is everything around that core. Raw materials differ by geography: sugarcane-derived feedstock for MSG in Japan, tapioca starch from cassava in Thailand. The chemistry is similar, but the agricultural and economic logic is local.

Product development also shifts with food culture. In Indonesia, for example, Ajinomoto Group sells seasoning products designed specifically for tempeh-based dishes, rather than simply exporting Japanese formats.

Distribution is another layer of localisation. Over decades in Southeast Asia, Ajinomoto Group has built relationships with small shop owners and foodservice operators and learned how to work across many sales routes, not just modern retail. That infrastructure is a real advantage for startups that need to adapt formats and pricing to fragmented, on-the-ground realities.

Even its startup investments follow this pattern. In Brazil, Ajinomoto Group backed Manioca, which develops products using tucupi, a traditional ingredient from the Amazon. The core ingredient and food science can travel, but the expression is rooted in local cuisine.

For founders, the implication is straightforward. Ajinomoto Group Ventures looks for globally relevant core technologies and strong ingredient R&D, paired with application layers that can be localised in product, branding, and route-to-market.

What startups must have before reaching out

For startups looking to reach out to Ajinomoto Group Ventures as a strategic investor, what should they have in place?

First, Ajinomoto Group Ventures usually does not lead rounds. The CVC takes a follower stance, with priority on strategic value rather than ownership targets or round construction. The CVC prefers to co-invest alongside an institutional VC that has already set terms and valuation. For Saito and his investment committee, third-party financial validation is an important part of the internal decision process.

Second, the strongest pitches come with a wide surface area for collaboration. A single pilot or narrow use case rarely justifies the effort of building a long-term relationship. Ajinomoto Group Ventures wants founders to lay out multiple ways the partnership could develop over time: R&D work, formulation and ingredient development, manufacturing, sales channels, and even future business hypotheses. Clear pathways give different teams inside the group something concrete to react to.

Third, founders need sharp competitive positioning. A clear explanation of who else is in the space and why the startup stands out. A crisp articulation of competitors and differentiation makes both collaboration and investment far more likely.

Finally, Saito emphasises that Ajinomoto Group Ventures is not a financial VC with a bit of “strategic” on the side. The group’s real value lies in collaboration, such as joint development, access to sales channels, and venture-client models, even when no equity is involved. Founders who approach the CVC with a partnership mindset, rather than as a funding source, tend to get much further.

A small ask from Saito

Saito is open to connecting with startups and investors aligned with Ajinomoto Group Ventures’ focus. If you are interested, feel free to reach out at hiroyuki.saito01.3cc@asv.ajinomoto.com.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Thanks for being here. For Issue #125 of Better Bioeconomy, I had the opportunity to interview Veronica Breckenridge, the Founder and Managing Partner of First Bight Ventures. First Bight is an early-stage venture fund positioned to capture the massive value-creation opportunity in the multi-trillion-dollar industrial transition from petroleum to bio-based manufacturing.

Veronica has one of the most interesting operating backgrounds I’ve come across in this space. She has scaled teams and complex supply chains at Apple and Tesla before moving into industrial biomanufacturing.

First Bight recently published the Industrial Biomanufacturing Graveyard report, a must-read report that turns decades of painful outcomes into decision-grade lessons. I open this article with a short primer on that report to clarify the context, then move into how Veronica applies those lessons in practice.

Let’s jump in!

The graveyard of industrial biomanufacturing is crowded. It includes biofuels that never cleared cost, platform molecules that ran out of runway, and plants that went far over budget.

None of that is a reason to retreat. It is a reason to learn. Failure leaves clues. It shows where market selection went wrong, where scale assumptions broke, and where capital stacks turned fragile under pressure. Read those clues well, and you avoid paying the same penalty twice.

That spirit is what inspired First Bight Ventures to publish its recent Industrial Biomanufacturing Graveyard report, one of the clearest attempts to turn decades of painful outcomes into a usable dataset and required reading for anyone building or backing industrial bio.

Veronica and her team treat the graveyard as a data source, not a warning sign. The result is a playbook that turns cautionary tales into operating procedures.

Key findings from the Industrial Biomanufacturing Graveyard report

The report maps where the first wave of companies aimed and where they fell. Most failures occurred after biofuels and broad “platform” molecules, which were low-margin, scale-driven markets where incumbents already sat on cheap feedstocks and fully depreciated assets. Only a small fraction started in higher-margin speciality applications, meaning many teams chose the hardest economic battleground as their first act.

Just as important as where they played is why they failed. Across the cases the report tracks, the dominant reasons are commercial and operational, not scientific. Market competition and poor product-market fit top the list, followed by scale-up issues and CAPEX overruns. Lab and pilot often worked well enough. The breakdown happened when processes hit the plant and could not produce at a cost or volume the market would accept.

Three recurring mistakes stand out. Founders consistently underestimated manufacturing cost at scale and assumed customers would pay a generous green premium that never materialised. They chased large commodity markets to make the TAM slide look exciting, only to collide with petro incumbents on price. And they led with novel products that required new supply chains and customer behaviour, adding friction at every step.

The report is equally sharp on scale-up and capital. First-of-a-kind plants frequently ran 3 to 5 times over budget because teams treated industrial assets like software, relying on equity-heavy cap tables and expecting smooth, rapid growth. Factories do not behave that way. They need staged risk removal, manufacturability designed in from day one, and capital stacks that blend equity, equipment finance, incentives, and strategic support. The failures often stemmed from capital that didn’t align with factory timelines, leading to premature scaling, unfocused plant builds, and a lack of discipline in techno-economic analysis.

The conclusion is not that industrial biomanufacturing is doomed. It’s that the game is winnable if teams internalise these patterns. Start in markets where biology can win on function and acceptable cost. Design for manufacturability and realistic price corridors from the start. Treat scale-up as a series of gates, not a single leap, and match capital to those gates with partners who know how to build and run plants. The graveyard is large, but it reads like a manual of what not to do, and a blueprint for doing it differently.

From Apple and Tesla to the future of industrial biomanufacturing

Veronica’s path runs through scaled execution before it runs through biology. At Motorola, she helped set the China strategy for the mobile devices business. She led a team scouting technologies that would push phones past voice into richer experiences, including an early prototype that foreshadowed what we now know as Siri.

At Apple, she built the Enterprise and Education business for Greater China, growing revenue nearly one hundredfold to about one billion dollars while building the teams, channels, and programs that turned a complex portfolio into repeatable sales.

At Tesla, she launched the company in China as Vice President, recruited and built the national organisation, opened stores across several cities, and owned the full stack from government relations to service and sales.

Those chapters were followed by a move into venture capital at Hone Capital, where she helped pioneer a data-driven seed strategy that produced one of the highest ratios of unicorns per assets under management. The through line is disciplined scale under real constraints, and the habit of turning frontier technology into systems that ship.

That operating muscle is what she now brings to industrial biomanufacturing. Her pivot came from a conviction she could not ignore: biology’s cost curves were accelerating and would eventually force a rewrite of how materials, chemicals, and ingredients are made.

She had seen this dynamic before. When cost curves bend, industries reorganise. Biology had the ingredients for a similar shift, but adoption kept stalling at the factory gate. Products missed price corridors, processes failed manufacturability gates, and materials did not slot cleanly into existing supply chains. That was the gap she chose to work on.

First Bight Ventures is her answer to that gap. She set it up to invest early and close the distance between lab, pilot, and purchase order. The firm pairs capital with hands-on de-risking and anchors that work through real buyer specs. It sits inside a regional buildout in Houston, with access to talent, partners, and pilot capacity that compresses learning cycles.

Why a concentrated portfolio works in industrial bio investing

Unlike a typical early-stage fund that backs 20 or more startups, Veronica runs a concentrated portfolio with just seven companies in First Bight. She shared that this is because industrial bio is manufacturing-heavy, and most early teams are scientific founders who have never run a plant or a business. They need help with commercial viability, manufacturability, market selection, and team buildout, which is hard to do from afar and demands a hands-on investor. A concentrated portfolio allows First Bight to provide that hands-on support to each of its portfolio companies.

The exit sizes of startups are another factor contributing to the concentration. Biomanufacturing is not software where you can rely on a 1000x outlier and a <1% hit rate to generate returns for the fund while all the other bets fail. Here, the realistic upside is more like 20-30x, so you need a concentrated portfolio with good survivability and exits, where one or two hits already get you to a 5-6x fund. To deliver that, you concentrate both ownership and attention.

Subscribe now

The founder profile that scales science and operations

Failures in biomanufacturing are rarely about science (or the market) alone. Many come down to the team’s ability to make disciplined choices as the company scales. That is why Veronica prioritises three signals when she backs founders.

Coachability comes first. Many scientific founders have not run a plant or a P&L. Founders who take feedback and adapt quickly are far more likely to grow into the operational demands of running a plant.

Business bias is the second signal. The founders she backs understand they are trying to commercialise a technology, not just build great science. Curiosity about customers, costs, and pricing separates teams that chase novelty from teams that build for adoption. She often finds this in repeat founders, including some who have failed before, because they have already learned the hard way what a real go-to-market and real unit economics look like.

The third signal is the founder’s ability to grow with the company. Companies change shape as they grow. The founder who can hire ahead of themselves or recruit a seasoned operator at the right moment improves survival odds.

Start with higher-margin niches with a drop-in pathway

Commercial traction shows up fastest when the product slots into existing infrastructure. Veronica prefers drop-in or near-drop-in molecules that leverage existing feedstocks, CDMOs, equipment, and downstream integration. Novel molecules can be powerful, but they often require changes at multiple points in the chain and long cycles of education. Those cycles typically belong on big-company balance sheets. Startups need proof sooner.

The Graveyard analysis backs this. Targeting commodities to tell a big story was common and pushed teams into scale traps they could not fund. The more feasible path, as Veronica emphasises, is to begin in speciality and higher-margin niches where functional advantages justify the price and volumes are small enough to learn without blowing out the cap table. Broader markets come later, once the process is validated and the cost story is real.

Defensibility still matters. Veronica is not interested in me-too processes that simply plug into existing systems. She wants technology that is meaningfully differentiated, but where adoption sits mostly inside today’s infrastructure rather than requiring a new system at every step.

This is less ideology than economics. Many founders assumed the market would pay a green premium. In reality, price corridors are tight, and procurement pressure erodes even modest premiums. Starting where biology offers a clear spec-level edge and can already run on existing lines buys time to fix yields, recovery, and downstream performance.

More radical, non-drop-in approaches will have their moment once biomanufacturing infrastructure is more built out and the industry reaches a more mature stage. For now, the priority is products that can clear real-world constraints and start flowing through the pipes we already have.

Equity capital efficiency beats the fantasy of CAPEX-light

When it comes to financing a first-of-a-kind (FOAK) facility, equity cannot and should not carry the entire plant. Veronica frames the goal as making equity capital efficient rather than pretending biomanufacturing can be CAPEX-light. FOAK means real hardware: fermentation tanks, utilities, downstream equipment, and all the supporting infrastructure.

Her push is to keep as much of that hardware as standard and financeable as possible. The more you rely on established equipment and standardised agreements, the easier it becomes to lease, refinance, or sell pieces of the asset if needed.

She takes the same view on government support. FOAK should lean on regional and national programmes that underwrite land, infrastructure, and jobs so that scarce equity does not pay for concrete and steel. Strategics play a similar role. Partners who have built plants before and carry large balance sheets can co-invest or hold parts of the asset, bring in their own engineers, and reduce execution risk in ways a startup alone cannot match.

The common thread is that the capital stack must align with the project’s physical reality. Venture equity is there to take early technical and commercial risk. Equipment finance, incentives, and strategic capital have to carry more of the load once foundations are being poured and lines are being installed.

Blueprint for building an industrial biotech company today

When I asked Veronica how she would build an industrial biotech company today, her answer built on themes we had already covered. She would still begin in a speciality, higher-margin niche with a clear drop-in or near-drop-in pathway, where the technology offers a functional edge and the economics can absorb early inefficiencies. The goal is a focused application that can reach profitability sooner and avoid the scale traps that pulled many first-wave companies into the graveyard.

From there, she turns to sequence and capabilities. She is blunt about bringing commercial leadership in early. B2B cycles are long, and selling into corporates is its own craft. Someone has to own the path from the first conversation to repeat order while the technical team runs experiments.

Corporate partners should also be engaged early so specifications, pricing, and use cases are shaped with a real buyer. Regulatory and quality expertise come in sooner than most founders expect, ensuring the process being designed will pass the reviews that matter.

Grants and other non-dilutive funding should target commercial milestones rather than pure science projects, so each tranche of capital is tied to progress that a customer can feel.

The manufacturing lead deserves the same early attention. Veronica wants someone who has commissioned lines and has lived through utilities, permitting, vendor management, and maintenance, all involved well before FOAK. That person will spot design choices that create headaches later and save quarters of runway by avoiding rework.

Near-term opportunities that biomanufacturing can bank on

For Veronica, the next investable opportunities in industrial biomanufacturing sit closer to the consumer, in categories that already support real margins. Personal care, food, and nutrition are at the top of her list, along with niche, high-end applications where speciality economics allow earlier profitability.

These categories are bankable now because they can absorb some early inefficiency, as long as the product matches petro-based incumbents on performance and lands in a price range buyers can live with. Feedstock strategy is part of that story. She prefers alternative or waste streams that lower COGS and avoid competing with food, which makes the whole proposition easier to finance.

Modular and local manufacturing is the long-run structural advantage Veronica sees for biology. Smaller, distributed assets that sit closer to feedstocks or customers and still hit spec are more attractive than a single mega-plant. She expects that shift to unfold over decades rather than years, with near-term wins coming from products that already match buyer priorities and can run on existing lines without major retooling. As she put it, “People forget the petrochemical industry had over a hundred years to reach today’s efficiency. Biomanufacturing just started.”

A small ask from Veronica

If you are an investor who is passionate about industrial biomanufacturing, already backing startups in this space, or looking to get more involved, Veronica and the First Bight team would love to connect. Reach out to contact@firstbight.com to explore how you might partner to support founders and help grow this emerging sector.

Takeaways from my conversation

• The graveyard is a dataset, not a warning sign. Treating past failures as structured information reveals patterns in market choice, manufacturability, and capital structure that teams can design around. The question shifts from “is this too risky?” to “which risks can we remove early?”

• Concentration and involvement beat portfolio volume. Biomanufacturing companies need investors who are close to market selection, manufacturability, team buildout, and scale-up decisions. Veronica’s venture model is a counterweight to the ‘more shots on goal’ approach, arguing that a tighter focus and deeper participation provide better returns.

• Market selection and product design must honour how the supply chain already works. Starting in higher-margin, drop-in or near-drop-in applications gives biology a fair fight. Adoption should live inside existing infrastructure.

• Forget “CAPEX-light”, optimise for equity capital efficiency. Biomanufacturing will always involve real hardware and construction risk. The job is to structure it so equity doesn’t carry the full weight.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now

Thanks for being here. For Issue #124 of Better Bioeconomy, I sat down with Mirasbek Kuterbekov, Principal at Thia Ventures, to talk about investing in the food-as-medicine space. Thia is an early-stage VC fund at the intersection of food, biotech, and health. The fund backs ingredients, enabling platforms, and clinical validation infrastructure rather than brand-only plays.

Mirasbek is a materiobiologist turned investor. He has shipped tissue-engineering work at Hoxton Farms, run product and process R&D at Dow, and published on cell-culture systems earlier in his career.

In this conversation, we cover why the last few years made food-as-medicine more attractive to invest in, what evidence earns a seed-stage check, where real moats live, the founder traits he backs, two pitfalls that still wreck startup pitches, and why “natural GLP-1s” are overhyped while maternal nutrition is underrated.

Let’s jump in!

Consumer agency is the unlock

Food and health have never been strangers. The missing piece has been proof you can use. For years, nutrition claims lived in marketing while the hard evidence sat in journals. Clinical readouts were slow. Manufacturing was expensive. Supply chains for novel ingredients were brittle. Too much good intent died in the gap between promise and purchase. But recent advances now let people feel and measure the effects at the consumer scale. The landscape has shifted from promises to readouts.

For Thia, the mandate was already in place. Over‑the‑counter CGMs (continuous glucose monitors) widened access. GLP‑1s gained broader indications and mainstream acceptance. Wearables reached scale with real feedback loops and raised the bar for products to show effect, not just slogans.

GLP‑1s made the shift tangible. As Mirasbek put it, the bigger trend is that they “give back people the agency.” Users report less “food noise” and find it easier to say no to dessert, which carries over into diet quality and other daily choices. The clinical story matters too. Weekly semaglutide cut major adverse cardiovascular events by about 20% in people with overweight or obesity but without diabetes, and US regulators now recognise cardiovascular benefit.

Wearables and consumer CGMs close the loop. As he noted, CGMs “can now be bought over the counter,” giving people “actionable insight into how their body responds to the consumption of glucose.” Dinner turns into next‑morning data you can act on: sleep score, time in range, resting heart rate. When a product helps, the numbers move soon enough for people to notice.

Agency compounds. Oura, a Finnish health technology company known for its Oura Ring that tracks sleep and physical activity, has member data showing that alcohol depresses heart‑rate variability and sleep quality, so users drink less once they see the impact. Fewer late drinks, better sleep, saner snacking, more consistent training. People feel it, see it in the numbers, and change.

For Mirasbek, this convergence of behavioural change and measurable outcomes shows why the space has matured. It signals a shift from potential to predictable value creation and marks a structural change. This feedback-driven behaviour makes the sector both more reliable and more scalable for investors seeking growth.

Evidence is the only antidote to hype

Food-as-medicine has become a buzzword lately, so how does an early-stage investor separate hype from real health outcomes when evaluating startups in this category? “Robust data in humans” is how Mirasbek draws the line between claims and credibility.

In a category drowning in wellness copy, he comes back to two sources of truth. First, draw on the strongest diet literature where it exists. Scientists broadly know which dietary patterns reduce risk; for example, the Mediterranean diet has demonstrated cardiovascular benefits in large, multicentre trials. The commercial challenge there isn’t to reinvent nutrition science, it’s to package adherence as convenience.

Second, fund genuinely new biology. Thia’s bar for preclinical novelty is animal studies. “You cannot just use in vitro models. You need to have tested on animals. At least some proof of efficacy in animal models makes the risk palatable.” If a startup can articulate a plausible mechanism, show efficacy beyond in vitro, and connect that story to a plan for human readouts, the risk becomes legible at the seed stage.

That early-stage discipline is also visible in Thia’s portfolio: the fund backs tools and platforms that professionalise food-health evidence, including People Science, which runs decentralised, cost-efficient human studies for nutrition and consumer health. For a sector that often outsources proof to marketing, a plug-in clinical engine is a force multiplier.

What are the moats in food-health?

It is one thing to offer a healthy meal or a supplement. It is another to have a durable, long-term competitive advantage. What turns a functional product into a defensible platform?

Founders often start with patents. Patents matter when the product rests on novel chemistry or a newly characterised metabolite, and Thia will back that. “You can have moats around novel ingredients and metabolites. You can patent those and create an enduring moat. But another one that often gets overlooked is branding, not flashy marketing, but consumer trust,” Mirasbek said. The quiet moats are often the most durable. Consistent, cost‑effective efficacy across lots wins over time.

Trust sits at the centre. It is built by dose fidelity and label accuracy across batches. Mirasbek cited a study that found 22 of 25 melatonin gummies deviated from label claims, ranging from 100% less to 350% more melatonin.

Exclusive supply is another advantage. Even in commodity‑heavy categories, specific functional ingredients can be chokepoints. “Access to specific ingredients and a resilient supply chain can be a moat. For example, protein-packed bar brand David bought its supplier to secure a key fat ingredient.” Exclusive or priority access to a critical input, whether a bioactive or a stabiliser, can lock competitors out of equivalent performance for years.

Subscribe now

Work with obsessive, evidence-led, and ethical founders

In seed-stage investing, the team is everything. In a previous interview, Mirasbek shared how “obsessiveness” about the problem and the niche is the top trait he looks for in a founder. Mirasbek looks for a particular kind of obsessiveness: the founder who is never done reframing the problem, never done testing the next version, never done asking whether the current approach is still the best one.

He pairs that mindset with two non-negotiables: evidence and ethics. The first keeps the company’s narrative honest as new data arrives. The second sets the culture. Transparent, rigorous, and safe to speak up. A simple gut‑check he uses while diligencing: would you want to work for these people? If the answer is no, something in the culture probably will not scale, and the startup will run into operational issues in the future.

Two avoidable pitfalls that wreck startup pitches

The first is regulatory blind spot. Food, supplement, device, and drug are not vibes. They are legal categories that set timelines, capital needs, and label language. “A specific disease claim changes the category. It comes with different costs and a different fundraising strategy,” he said.

Cross that line and you are no longer building food as health. You have become a biotech or device play with a very different road map and investor set. Get the claim taxonomy right early. For example: “supports healthy blood sugar” is a structure or function claim; “prevents type 2 diabetes” is a disease claim.

The second is overreliance on borrowed science. Citing broad nutrition papers is not a moat. “It is not wrong, but it does not distinguish you. Having a specific ingredient like fibre does not give you a moat unless it is unique or validated through your own trials.” If your effect depends on a particular isomer, dose, or matrix, show it. Run your own protocol or use a credible platform to generate human readouts tied to your product.

Natural GLP‑1 are overhyped, and maternal nutrition is underrated

“Overhyped? Natural GLP‑1s. They get under my skin. There is some science there, but the idea that food can compete with medicines from Novo Nordisk and Eli Lilly is not true,” Mirasbek shared.

Biology explains why. Endogenous GLP‑1 is produced in small amounts and is rapidly degraded by the enzyme DPP‑4. Its half‑life is one to two minutes. Long‑acting analogs are engineered for albumin binding or DPP‑4 resistance and reach half‑lives measured in days. The credible opportunity sits beside the drug. Build foods and supplements that reduce side effects, improve tolerability, and help people stay on therapy.

What’s underrated? Maternal nutrition. Growing a human creates different nutritional demands, yet research and product support remain limited. The brief is clear: keep taste and convenience high, and sugar low. Make it easy for expecting mothers to eat in ways that help them now and give their children a better start.

Mirasbek cited evidence that points in the same direction. Early‑life sugar exposure and maternal metabolic state matter for long‑term cardiometabolic risk. Recent analyses of UK cohorts exposed to post‑war sugar rationing found lower adult cardiovascular events among those who experienced low‑sugar environments in utero and infancy.

A decade of work in infant‑focused functional ingredients strengthens the case. From 2′‑fucosyllactose to broader HMO mixes with plausible immune and gut benefits, the science is moving from safety toward outcomes that matter to parents and paediatricians.

If you want a food‑health domain with high evidence standards and clear long‑term value, Mirasbek believes maternal and infant nutrition is a compelling place to build.

A small ask from Mirasbek

Are you a founder or investor interested in food as health? Feel free to reach out to Mirasbek at mirasbek@thia.ventures.

Takeaways from my conversation

1. Agency and feedback loops drive investability: GLP-1s quiet “food noise” while over-the-counter CGMs and wearables turn meals into next-morning readouts. Together, they widen the user base and raise the bar for products to show effect. It is no longer just about health potential. Measurable behaviour change is now something investors can underwrite.

2. The bar for seed-stage is human-oriented evidence, animal before people: Two valid paths earn a check. Founders can build on large, public diet evidence where it already exists, or bring new biology with a plausible mechanism and animal efficacy on the way to human readouts. In vitro work alone typically does not make the risk legible at the seed stage.

3. Moats come from trust and supply, not just technology: Consumer trust compounds through label accuracy and consistent dose delivery. Over time, that reliability becomes brand equity. Controlling a key ingredient or securing a resilient supply chain is another practical moat in categories that often look like commodities from the outside.

4. Regulatory map before product map: Food, supplement, device, and drug are legal categories that define timelines, capital needs, and label language. A single disease claim can move a company into a different category with higher costs and different investors. Understand the claim taxonomy early.

If you found value in this newsletter, consider sharing it with a friend who might benefit from it! Or, if someone forwarded this to you, consider subscribing.

Subscribe now