How does Horizon fund different types of biotech?

Introduction

Horizon Europe grants to private firms offer a useful window into where EU public R&D money is actually landing, and within biotech, health and non-health firms follow markedly different patterns. This Brief focuses on that split. We leaned on the most authoritative definition available: the European Commission’s proposed Biotech Act, which distinguishes explicitly between health and non-health biotech applications. A machine learning classifier trained on expert-labelled project descriptions and the Act’s own definitions does the sorting. The result is two firm profiles that look quite different from each other.

Industrial biotechnology accounts for the largest share of firm non-health biotech projects in the Horizon dataset (approximately 2/3rds of funding). In practice, this means engineered microorganisms used to produce chemicals, materials, fuels, food ingredients, and agricultural inputs from feedstocks (e.g. sugars, agricultural residues, industrial CO₂, and waste streams). How this works is scientists modify the DNA of a microorganism, usually a bacterium or yeast, so that it produces a specific molecule of interest in a predictable and optimised way. The modified cells are grown and then transferred into fermenters (large steel tanks) where they are fed a feedstock and held at the right conditions to produce the target molecule: a plastic precursor, a flavouring, a fuel, a fertiliser ingredient. Industrial biotech sits at the heart of Europe’s bioeconomy strategy and its ambitions for strategic autonomy. And it is unlike any manufacturing technology that came before it: conventional industrial processes use chemistry and predictable physics, while biomanufacturing uses living cells that respond, adapt, and evolve. In essence, the cell is the ‘factory’. Fermentation itself is ancient, but engineering cells to have improved characteristics (eg. higher yield, quality, robustness) and/or to make non-native molecules, and harnessing them efficiently and reliably at industrial scale, is something we are still learning how to do.

The premise behind funding industrial biomanufacturing is that bio-based production can substitute for current petrochemical processes producing most of our everyday materials, chemicals, and fuels. With bio-based processes facing a structural cost disadvantage against 150 years of petrochemical infrastructure, supply chains, and input prices that have never incorporated their environmental costs – strategic public investment to facilitate industrial biotech’s cost competitiveness is critical.

This brief asks if Horizon funding is reaching the firms that would build that robust future, or reinforcing what Europe already does well, incrementally? Commercialisation is constrained by many things at once, with no single one explaining why so few processes reach the market: feedstock costs, scarce scale-up infrastructure, weak market pull against amortised petrochemical incumbents, and thin late and/or early-stage capital and talent. This brief follows the public money, asking which firms and stages the grants reach, and what happens once the firms are funded.

What does the data show?

Finding 1: The typical non-health biotech grant recipient is older, larger, and less risky than its health counterpart

The median health biotech firm receiving Horizon funding is 10 years old, employs 12 people, and turns over around €1.1M a year. One in four is still loss-making three years after the grant comes in – in other words, very small, early-stage companies, operating in a market that hasn’t paid off yet.

Non-health biotech firms look quite different. The median recipient is older (15 years), larger (23 employees), and generates four times the revenue compared to its health counterparts – around €4.3M a year. Fewer than one in five is loss-making after the grant. These are small but established businesses, with a lower-risk profile and a foothold in relatively niche markets. 

Finding 2: Funding concentrates in growth-stage firms in non-health biotech and relatively underfunds early-stage firms

We sort firms into early-stage, growth, and mature tiers using an equal-weighted score across headcount, age, and revenue, with cut-offs aligned to the EU’s SME definition. We classified 70.7% of biotech firms into a stage.

In non-health biotech, growth-stage firms take the largest share of grant money: 50.9%, or €231M. Early-stage firms make up a third of recipients (32.6%) but receive less than a quarter of the money (22.9%, or €104M). Mature firms take the rest (26.2%, or €119M).

Health biotech distributes differently. Early-stage firms make up 44.8% of recipients and receive 37.1% of the grant money (€485M). As the Sankey flow charts show, early-stage health firms draw roughly €139M in the €2M-5M bracket and €24M above €5M, compared with only about €19M and €0M, respectively, for early-stage non-health firms in those same brackets. These large grants to early-stage health companies include EIC Accelerator awards of around €2.5M, routinely going to high-growth young firms.

This gap may partly reflect fewer early-stage non-health firms existing and/or applying to Horizon, rather than allocation alone. In any case, this thin early-stage pipeline is a signal of an undersupported layer of the ecosystem, one which the funding system should seek to strengthen given the EU’s large ambitions for the sector.

In non-health biotech, the grant distribution is highly concentrated in growth firms – is Horizon funding what Europe already does well, or what Europe needs to build?

Finding 3: Horizon funds what Europe already does well rather than category-creating firms

Non-health biotech, fermentation, agricultural applications, and bio-based chemicals, have been part of Europe’s industrial fabric for decades. The EU bioeconomy generated €863 billion in value added in 2023, employing 17.1 million people; biomanufacturing has grown at 4.7% annually since 2008, outpacing ICT, transport, and financial services^[1]–^[2]. Scaling bio-based processes is genuinely capital-intensive. The instruments directing the largest tickets, BBI-JTI under Horizon 2020 and its successor CBE JU under Horizon Europe, are public-private partnerships designed specifically for this: funding the scaling and demonstration of bio-based processes, moving them from pilot stage into full commercial production. They were built for growth-stage firms, and the grant distribution reflects that.

What is harder to explain is that mature incumbents absorb 26.2% of grant value (€119M), and early-stage firms are underrepresented at 32.6% of recipients that receive only 22.9% of grant value, compared to 44.8% and 37.1% in health biotech. More than four in five early-stage non-health firms (211 of 255, 82.7%) never receive a grant above €500k. A grant of that size is unlikely to fund meaningful technology development in biotech. A 2023 McKinsey study found that Europe already leads the US on incremental innovation patents in food technology, macromolecular chemistry, and environmental applications^[3]. However, at the frontier (the authors singled out cell printing and engineered bacteria specifically^[4]) US players hold nine of ten top patent positions and investment runs four times higher. The authors identified the absence of bio-first platform companies, firms that treat rewiring biology itself as their core technology, as Europe’s most notable structural gap. Most of these innovative firms are, especially in Europe, early-stage^[5] – they are creating new categories rather than scaling established ones. Horizon funding in non-health biotech is thinnest in exactly this segment. This maps onto Bocconi’s broader analysis of Horizon using our same database: Europe tends to fund middle technologies and what industry already does well rather than disruptive innovation.^[6]

Finding 4: the technology base is shifting toward bio-digital approaches, but funding hasn’t fully followed

The technological underpinnings of non-health biotech are shifting. Roughly 70% of non-health Horizon funding flows into food, agriculture, bioenergy, and bio-based chemical applications – established domains with strong European industrial roots. Digital and AI approaches are present and growing in non-health biotech, from around 12% of projects in H2020 (2014-2020) to 23% in Horizon Europe (2020–2027), though projects remain largely product-specific: optimising one enzyme, modelling one pathway, improving one fermentation process.

What comes next in biomanufacturing, bio-digital manufacturing, sits at the convergence of two fields. On the biological side, engineered organisms and redesigned metabolic pathways are expanding what living systems can produce. On the digital side, AI models learn from each run, predicting how a process will behave, flagging when it drifts, and adjusting it in real time, so that output becomes more reliable and cheaper with every cycle.

Its building blocks are synthetic biology, AI and machine learning for designing and optimising bioprocesses, and digital platforms for running them. These are general-purpose capabilities. Together they change how biological production is designed, tested, and scaled across food, materials, chemicals, agriculture, and energy alike.

Competitive advantage in the next wave of non-health biotech will belong to those who fund these capabilities as shared infrastructure that many sectors can build on, rather than as one-off applications tied to a single product.

Whether established players are building these capabilities, or positioned to, is an open question. A healthy bioeconomy innovation ecosystem needs grants reaching both the established firms scaling proven processes and the frontier players experimenting outside the box,^[17] building the synthetic biology, AI-driven bioprocess design, and biomanufacturing platforms the next generation of European bio-based industry will depend on.

Finding 5: The Horizon science-firm innovation relationship is weaker in non-health biotech

Across three separate measures, non-health biotech sits much further from fundamental science than health biotech does. Just 7.6% of non-health biotech Horizon funding flows through Pillar I, the science excellence pillar where frontier and early-stage research happens, against 20.7% for health biotech. Only 26.4% of non-health biotech projects are at low TRL (1–3), against 43.1% for health, and universities hold 42.1% of consortia places, against 50%.

Non-health biotech competes on the market on different terms than health biotech does. A new drug can succeed on almost efficacy alone compared to existing treatments, relatively independently of production scale. In non-health domains, a bio-based industrial process producing, for example, a chemical or material competes on margin against a petrochemical process that has often had decades to pay off its capital investment and optimise its economics.

This competition happens at an industrial scale, so reaching that scale is a precondition for competing with incumbents. Scale is necessary, but not always sufficient: margin depends on factors beyond plant size including feedstock availability and price, the chemistry of the target product, the capital cost of a plant, the operational cost of a plant, secondary processing steps of the product, and the ability to produce consistently at volume. 

For some products the gap is structural, when the feedstock costs nearly as much as the finished product, even a highly efficient process cannot reach parity, and scale or efficiency gains do not change that. However, where the economics are close enough to compete, how efficiently the biology converts feedstock into product often decides whether a process succeeds, and that efficiency is a question of fundamental science.

We find that non-health biotech features less in Pillar I funding and low-TRL work. This points to less capacity to do exactly the kind of fundamental synthetic biology research which could lead to increased robustness and performance of the organisms doing the work. This is further expanded in the section below.

Finding 6: The grant effect on revenue differs between health and non-health and by stage in biotech

The descriptive picture above shows that health and non-health biotech receive Horizon funding through quite different distribution patterns, with different firm profiles, different ticket sizes, and different stages of firm development absorbing the money. The next question is whether Horizon funding to firms actually translates into revenue growth, and for whom.

We draw on the IEP-COMPET dataset^[18], linking CORDIS project records to Orbis financial accounts. We restrict to firms with a single Horizon grant, controlling for firm and year fixed effects, to isolate the grant’s effect on revenue. The biotech sub-sample is small once cut by sub-sector and stage, so these findings are directional rather than definitive.

Early-stage health biotech shows a significant positive effect of the grant on firm revenue outcome. European health biotech sits on dense pre-existing research clusters, a specialist investor base familiar with its development pathway, and an EIC instrument suite that routinely directs grants of around €2.5 million to high-growth early-stage firms. Horizon appears to be reinforcing an ecosystem that already absorbs public investment productively.

Early-stage non-health biotech, by contrast, shows no significant revenue effect. To note, the revenue coefficient is larger than for growth-stage non-health biotech firms, so the lack of significance could be attributed to high variability in a small sample size.

The descriptive picture in Finding 2 already shows that this layer of the ecosystem is underfunded relative to its health counterpart, with early-stage non-health firms receiving 22.9% of grant value against 37.1% for their health equivalents. What the regression adds is that even when these firms do receive grants, the grants do not translate into measurable revenue growth. More than four in five early-stage non-health firms receive grants below €500k, and the regression suggests that those small tickets do not effectively allow early-stage firms to develop. What may be missing is not only large enough grants, but the wider ecosystem that translates capital into growth. This could be infrastructure and scale-up gaps or the absence of the kind of mature, specialist investor base that early-stage health biotech can draw on, but that is lacking for non-health biotech.

The growth-stage results reverse the pattern. Non-health biotech at the growth-stage responds positively to grant receipt (p<0.05), which is consistent with the design logic of CBE JU and BBI-JTI: once a firm has cleared early process validation, the remaining constraints (yield, downstream processing, capacity) respond to incremental capital. Health biotech at the growth-stage shows no significant effect, most likely because a large share of these firms are running clinical trials whose binary outcomes decide success or failure regardless of grant funding.

For non-health biotech, two findings stand out.

1. The growth-stage positive result validates the existing scaling architecture. Funding to growth-stage non-health firms has a positive effect on revenue, so continued scale-up support has empirical backing rather than resting on assumption alone.
2. Early-stage firms are where disruptive innovation tends to happen, and where a grant is smallest relative to firm size, so the revenue effect should be largest here if anywhere. Instead it’s absent. That points to an ecosystem not yet configured to support disruptive non-health biotechnologies, which is a troubling finding given Europe’s ambitions in the sector.

Growth-stage funding is working on its own terms. The question is what it will have to work with in ten years, given that it is fed by an early-stage layer being underfunded today.

What this means for policymakers

An incremental bioeconomy – by design?

This data is consistent with a Horizon portfolio weighted more toward incremental work in the niches where bio-based manufacturing has already commercialised than toward the pioneering platform development that would robustly expand the sector. Part of this tilt is structural to how the relevant instruments are programmed: The Joint Undertaking model that channels the largest non-health biotech tickets (CBE JU and its predecessor BBI-JTI) gives industry partners deep influence over the research agenda (according to some commentators, this includes over project evaluation), a design that critical analyses argue reinforces incumbent priorities rather than opening space for frontier and disruptive entrants.^[19] In non-health biotech, grants concentrate in growth-stage, more established firms, while early-stage firms receive a smaller share and show no measurable revenue effect from the grants they do get.

This does not mean frontier work is absent from European non-health biotech. The convergence of synthetic biology and novel metabolic engineering with machine learning and predictive digital twins represents a qualitative shift in what bio-based production can do and how the Valley of Death between lab demonstration and commercial viability can be crossed. This frontier work is happening in European science, for example at DTU Biosustain, Wageningen, and Chalmers, but the early-stage companies that should be translating this science into commercial development are underrepresented in Horizon’s non-health biotech portfolio. The lower university share in non-health biotech consortia (42.11% against 50% in health) hints at exactly the structural disconnect identified in research on the valley of death: academic discovery and industrial application operating at different speeds, with weaker bridging mechanisms to enable disruptive innovators than comparable sectors.^[20] – ^[21]

We should continue funding scaling and demonstration – that work needs to expand and to be coherently designed to have a create the full bio-stack across CBE JU, BBI-JTI, InvestEU, IPCEI, and the emerging tools of the Biotech Act. But Europe also needs to nurture the innovation pipeline scaling depends on. Early-stage firms pursuing disruptive innovation sit at a stage where private investment is typically harder to secure given the technical and regulatory risk. In non-health biotech, this early-stage layer receives a fraction of what health biotech does. This is the pipeline that eventually produces the technologies that scale; without it, demonstration funding has, quite literally, less to demonstrate. The Biotech Act has an opportunity to develop this, particularly the enabling layer of cross-cutting tools, platform logic and digital tools for enhanced scale-up behaviour prediction, alongside its scale-up provisions.

Competing at scale: constraints are systemic

Our results show that non-health biotech features less in Pillar I funding and low-TRL work. This points to less capacity to do the kind of fundamental synthetic biology research which could lead to increased robustness and performance of the organisms doing the work. The following section looks at why underfunding early-stage research-adjacent projects matters. We explore titre, rate, and yield (TRY) as an example of a cost constraint rooted in fundamental science, then turn to scale-up and pilot plants where lab research is tested against industrial reality. Lastly, we speak to federated data as the missing link between lab and plant

Three intertwined processing metrics – TRY, can constrain the cost competitiveness of a new bio-based process.^[22] Techno-economic analyses across precision fermentation, bio-based chemicals, and one-carbon biomanufacturing consistently identify low TRY as the binding constraint on cost competitiveness, because poor yields force larger fermenters, higher capital expenditure, and inefficient downstream processing, which produce order-of-magnitude cost differences between a lab-scale demonstration and commercially viable production at industrial scale.^[23] Research on the ‘valley of death’ has identified this TRY gap as fundamental, finding that biotech commercialisation remains structurally blocked at the science-to-industry interface.^[24]

Improving TRY depends on fundamental synthetic biology research at every level of the cell.^[25] At the smallest scale, that means engineering individual enzymes and the chemical pathways they drive. At the largest, it means mapping and redesigning a cell’s entire metabolism so the whole organism works as a dependable production factory. A 2020 study shows what this looks like in practice: A genome-scale metabolic model was used to work out, at the genetic level, which reactions to remove so that the cell would make the target product chemical as it grew. In a single Design-Build-Test-Learn (DBTL) cycle the redesigned strain reached 50% of its theoretical maximum yield, and it held that performance all the way from small shake flasks to industrial-scale bioreactors.^[26] The researchers estimated that the same model could reach 90% of maximum yield over further DBTL cycles. The gap between what has been demonstrated and what is possible is fundamental science still to be done, and doing can be decisive on whether a process is cost-competitive and robust at industrial scale. Bio-digital approaches are increasingly how that work happens, combining AI-driven strain design, machine learning for predictive metabolic modelling, digital twins for fermentation, and adaptive laboratory evolution, all embedded in the fully integrated DBTL cycles that apply the scientific method through prediction and iterative learning.^[27]

Industry has every interest in better TRY. However, the fundamental science that produces step-change improvements is rarely the kind of research and development that private firms fund, because the returns are diffuse, the timelines are long, and the science sits too far upstream. Work of this kind depends on public investment, and supporting it is what Pillar I exists to do.

The science capable of fundamentally shifting what TRY can achieve, rather than incrementally improving it, is the kind Pillar I is designed to fund: work with diffuse returns and long timelines that private firms are least likely to fund on their own. However, non-health draws barely a third of health’s Pillar I share of funding, and only 26.39% of its projects sit at low TRL against health’s 43.07%. Yet this is the work that can determine whether a bio-based process can undercut a petrochemical one, and whether it survives the jump from a few litres in a flask to tens of thousands in a reactor. If bio-based production is going to close the cost gap with petrochemical incumbents, the science that can help get it there, and keeps it robust at scale, needs a Pillar I allocation proportional to the science it requires, not a fraction of it.

Scale-up and pilot plants: where the science-commercialisation relationship becomes real

Much of the conversation about bio-based scale-up centres on physical capacity and its cost: large-scale fermentation tanks, downstream processing facilities, and the engineers who run them, all of it capital-heavy and slow to build. This is where the well-documented valley of death sits,^[28] in assets that are expensive, intermittently used, and unglamorous, generating none of the intellectual property a novel organism does.

A pilot plant is a physical site, but it does something more specific than housing equipment: it is where lab-stage biology and industrial engineering problem-solve together, testing whether upstream science actually holds up once commercial conditions apply. This convening space, where research, scale-up science, and commercial realities meet, can start to close the science-firm gap the Horizon data points towards.

Currently, a big constraint is that this infrastructure has to be built rather than rented. This goes against the classical biotech investor logic as they were largely trained in pharma, where a mature contract manufacturing network can be plugged in late. That model does not transfer to food ingredients or bio-based chemicals, where margins are thinner, volumes are orders of magnitude higher, and platforms are usually bespoke to a single process. No one firm can justify owning capacity it uses only intermittently, which makes the lack of shared pilot infrastructure a market failure that public investment exists to solve.^[29] The US has committed over $200M to shared facilities at 5,000–25,000 litre scale on exactly this logic. Europe’s main instruments, the BBI JU and its €2 billion successor the CBE JU, instead de-risk deployment plant by plant, so the asset stays private and does not compound.

Brief I sets out what the alternative looks like: a coordinated European pilot plant network built on three compounding layers, shared physical infrastructure for scale-up, a federated data layer that turns every campaign into transferable process intelligence, and the training infrastructure that supplies the operators and engineers to run it. These are the spaces where foundational research and commercialisation actually meet, where a result from the lab is tested against industrial reality and either holds or breaks. In essence, researchers are engineering tiny living factories, so what fails at scale feeds back into the basic science, telling researchers which organisms, pathways, and processes need rethinking. A pilot plant network is therefore not only a bridge from science to market but the mechanism that keeps the science itself improving to meet real world conditions and commercial considerations.

Visible only at scale: science, infrastructure, and the cost gap

Adequate infrastructure and trained operators is the prerequisite for scale-up; the science is what then has to succeed on it, and the hardest scientific challenges surface only once a process reaches industrial scale. Post-mortems of the industrial biotech commercial failures find the same recurring causes: organisms that perform well in a two-litre flask behave differently in a 100,000-litre fermenter; microbes that lose their engineered properties over a production run as natural selection works against the modifications; processing costs that fail to fall with volume the way financial models assumed.^[30] KiOR (failed biofuel company) projected 67 gallons of fuel per ton of biomass in its investor filings; actual production yielded far less.^[31] Amyris (which converted cane sugar into a chemical precursor using engineered yeast), hit 15% yield at lab scale and couldn’t replicate it in production with the company eventually filing for bankruptcy.^[32]

For products competing head-on with petrochemicals, feedstock economics can be decisive, since even a perfect fermentation cannot close the cost gap when the raw material costs nearly as much as the finished product. This does not weaken the case for better science, instead it pushes foundational research further upstream, toward alternative feedstocks, waste-derived substrates, and CO₂-based conversion routes where fermentation gains can then be realised.

No single bottleneck explains the pattern of commercial failures in industrial-scale biomanufacturing, because the causes are systemic. Feedstock costs can make a process uneconomic regardless of how well the biology performs, particularly for low-added products. Downstream processing can consume more capital than the fermentation itself. And building a first-of-a-kind production plant is itself a structural disadvantage: the financing costs alone make it nearly impossible to compete on price with petrochemical facilities that were built decades ago and have long since paid off their debts. TRY, the efficiency with which an engineered organism converts inputs into outputs under real production conditions,^[33] is not always the first constraint to bind, but it cuts across all the others: poor fermentation performance forces larger reactors, drives up downstream costs, and amplifies feedstock waste. It is also the constraint that most often reveals itself only at industrial scale, when living systems meet conditions that cannot be replicated in the lab. All of this decides one thing: whether a bio-based process can undercut the petrochemical one it aims to replace.

Closing the loop: federated data as the missing link between lab and plant

Closing that gap increasingly depends on something the industry, as a collective, does not yet have enough of – data. Predicting how a strain behaves over a full industrial production campaign requires federated bioprocessing intelligence, meaning shared cross-process data that makes real prediction possible.^[34]

Building that predictive capability is cross-cutting, platform-level work that a single firm will not fund, because a single firm cannot capture the returns. This is one part of the infrastructure coordination our piece Towards a European BioPower called for – a network of cloud labs and European biofoundries, built as shared infrastructure to democratise advanced tools and let SMEs prototype at scale. This kind of shared infrastructure extends past research to scaleup: pilot plants that generate bioprocessing data that can feed back into the fundamental synthetic biology research to produce industrially-robust strains and back into process runs. The data layer comes from a concerted effort to identify which forms of data to collect and share at the European level, and how to integrate process improvement learnings. Recent research points to what this investment should look like in practice:^[35] 

• Federated data-sharing systems that allow companies to pool operational knowledge from industrial campaigns without surrendering commercial secrets;
• AI models trained not just on laboratory data but on the full complexity of industrial fermentation, including mixing dynamics, stress responses, and genetic drift;
• Digital twin platforms that can simulate a scale-up campaign.
• data standards to support the building of foundational models (read more in our Call For Evidence for Biotech Act II)

Advanced biomanufacturing is the science of living factories, systems of great biological complexity, designed by evolution over billions of years, now being re-engineered to work for us. Understanding them well enough to make them reliable at industrial scale, with the help of data and learning systems, is one of the defining scientific challenges of the next decade. It is exactly the kind of early-stage, infrastructure-enabling science that public investment exists to support, and that the Biotech Act II has an opportunity to make deliberate space for, alongside its scale-up provisions.

The bioeconomy cannot be granted into existence 

Horizon grants can take non-health biotech from lab bench to pilot facility and to commercialisation demonstration, but even a well-calibrated programme cannot resolve the deeper market failure at the heart of bio-based commercialisation. Demonstration and scale-up instruments do two things at once: they prove that bio-based alternatives actually work, and they stimulate the bio-based industrial substitution that is critical for Europe’s strategic autonomy and resilience. But they fund one application at a time, and one application at a time cannot build a full bioeconomy. Bio-based processes face a structural cost disadvantage against 150 years of petrochemical infrastructure, supply chains, and input prices that have never incorporated their environmental costs – and this was not simply the result of market forces, it was actively constructed. The Commission acknowledges as much: “the current oil-based economy was built through a century of economies of scale” and that “similar efforts are required to build this capacity for bio-based products”.^[36] The US federal government introduced tax breaks for oil and gas as early as 1916; over the subsequent decades, federal subsidies to the industry averaged $1.8 billion per year in inflation-adjusted terms. European governments pursued comparable industrial strategies through preferential tax treatment, state ownership of national oil companies, and infrastructure investment. The bio-based sector has received nothing remotely comparable. Public grants and equity investment can de-risk a bio-based process enough to attract follow-on private investment, but neither changes the cost structure the resulting product meets on entering a market shaped by a century of subsidy to the petrochemical incumbent it aims to replace.

The main lever to tackle this sits outside Horizon’s remit: market-shaping policy that prices in environmental costs and creates durable demand for bio-based products. The EU has the tools to do this. Without using them at a meaningful scale, even successful demonstrations will remain perpetually subsidy-dependent.

Non-health biotech needs four interventions working together:

1. Continued scale-up support.
2. Market conditions that pull bio-based products into a competitive position.
3. An innovation pipeline reaching early-stage firms and frontier science to produce what the market can then pull.
4. Cross-cutting bio-digital capabilities that support and accelerate learnings from science to successful commercialisation

Conclusions

Findings: Non-health biotech firms are older, larger, and less risky than health biotech firms, and Horizon funding to them concentrates in growth-stage, more established players. Early-stage non-health firms receive a smaller funding share than their health counterparts and, unlike early-stage health firms, show no measurable revenue effect from the grants they do get. The lack of post-grant revenue growth for early-stage non-health biotech firms points towards a wider ecosystem gap (weaker infrastructure, thinner specialist investment), but the sample is too small to confirm it. Non-health biotech also sits further from fundamental science on every measure: lower fundamental science funding, lower low-TRL share, lower university involvement in projects. Growth-stage non-health firms do respond positively to funding, validating the existing scale-up architecture.

Conclusion: Non-health biotech funding rewards what Europe already does well rather than building the cutting-edge frontier (bio-first platform companies, bio-digital manufacturing) where Europe is weakest against the US.

Suggestions: Rebalance funding toward early-stage, science-adjacent non-health biotech. Pair continued scale-up support with market-shaping policy that prices in bio-based products’ environmental advantage, since grants alone cannot offset a century of fossil-fuel infrastructure subsidy and scale.