Does Horizon funding fit the biotech firms it's trying to grow?

Introduction

You may have heard that Europe is having its biotech moment. The proposed Biotech Act (2025) aims to accelerate and facilitate how biotech products move from lab to market, and to unlock investment for biotech companies in Europe, including a €10 billion investment pilot run jointly with the European Investment Bank. It is arriving in two parts: an initial package on health biotech, followed by a second package on non-health biotech. Separately, the EU institutions are negotiating the bloc’s next seven-year budget (the 2028–2034 Multiannual Financial Framework), which will determine funding for research through the next Framework Programme, Horizon Europe (informally FP10), and for industrial competitiveness through a new and closely linked European Competitiveness Fund.

All three efforts agree on the same premise: biotech is strategically essential, private capital alone will not build it, and public money has a role to play in bridging the gap from science to commercial production. Biotech is a sector where public funding can be instrumental as risk is high, regulatory timelines are long, capital requirements are heavy, and the pre-commercial phase stretches well beyond what markets will easily fund on their own.

This brief asks whether the public money currently doing most of that bridging reaches the firms the strategy depends on, in the form they actually need. It compares EU funding to biotech firms against funding to their peers in two groups we define for this analysis. Digital tech covers software, AI and computing; hard tech covers hardware-based sectors such as semiconductors, advanced materials, and clean energy. The groupings follow the EU’s own technology categories (STEP, the EIC, the Key Enabling Technologies), and firms are sorted using project descriptions, NACE codes and patent data. We looked only at firms that won Horizon funding, the firms the system already selected for. If the instruments fit even these firms poorly, they are unlikely to fit unfunded firms any better. Other instruments, including InvestEU, structural funds, and national programmes, fall outside the scope of this analysis.

Brief I considers three questions:

• What does the typical Horizon biotech firm actually look like, and how does it compare to its peers in digital and hard tech?
• What does the financial data say happens to those firms once they receive a grant?
• And what do the patterns suggest about whether Europe’s funding architecture fits the firms it is trying to grow?

What does the data show?

Finding 1: Digital tech and biotech – similar firm profiles, differing development needs

On standard measures, biotech firms receiving Horizon funding look statistically like their digital counterparts. Both sectors skew young: 44% of biotech firms and 48% of digital firms are under ten years old, against just 38% in hard tech. Both run lean, with median headcounts of 15 and 17 employees respectively, compared to 25 in hard tech. And both generate modest revenues – €1.8 million median in biotech, €2.3 million in digital – well below the €5.0 million median in hard tech. Hard tech sits clearly apart on every count: older firms, larger workforces, and a much heavier upper tail running into the hundreds of employees, as you would expect from sectors built around physical hardware, specialised infrastructure, and often long development cycles.

Despite the similarity between biotech and digital tech firms participating in Horizon, the underlying development process for biotech is categorically different to digital tech: It runs on long regulatory timelines, heavy capital expenditure, physical hardware, and laboratory to scale-up infrastructure, with a slow path from scientific proof-of-concept to commercial product^[1]. On those terms, biotech’s development logic sits much closer to hard tech. Biotech firms are combining the development demands of hard tech with the firm scale and balance sheets of digital tech^[2]. This is a structural tension at the heart of biotech’s commercialisation challenges.

Finding 2: Biotech firms skew towards earlier stages of technology readiness development

Firms participating in Horizon biotech projects show a distinct orientation towards earlier-stage, research-intensive activity than in comparable deep tech sectors. These are private, for-profit companies whose R&D is by definition commercially oriented, yet even this commercially relevant research remains tied closely to academic institutions.

Unlike in digital tech, where a firm draws on research to build a product and then largely departs from it, biotech development tends to stay scientifically intensive well into commercialisation. Scaling a biological system, optimising a fermentation process, qualifying an engineered cell strain for consistent production at volume – these involve genuine scientific and technical uncertainty when moving from lab to facility. Scale-up is increasingly recognised as a primary technical bottleneck: processes that perform reliably at bench scale can behave unpredictably at industrial volume, requiring original problem-solving at each transition point^[3].

The broader literature makes the point that biological innovation isn’t cleanly sequential as research to engineering to market do not hand off in tidy stages^[4],^[5]. It’s more of a continuum, where scientific and technical challenges are interwoven throughout the development cycle. As a result, policies and instruments that delineate development stages and scientific challenges strictly tend to impede technology commercialisation.

The data reflects this: forty percent of Horizon biotech projects are classified as being at low technology readiness levels (TRL) (1-4), more than double the rate of digital (13%) and hard tech (16%). The median biotech project has a university partner share of 50%, well above both digital and hard tech at 36% each, and biotech firms receive 17% of their EU funding through Horizon’s basic science pillar, compared to around 5% for both digital and hard tech.

Universities contribute well beyond early-stage discovery into scaling lab results to industrial production, optimising bacterial strains, and the analytical testing that confirms purity and quality^[6]. For smaller biotech firms, having academic partners allows for access to containment facilities, bioreactor systems, and analytical equipment that is prohibitively expensive to own independently. This relationship is in many ways symbiotic, a natural extension of the technology translation and knowledge-transfer roles that universities have come to embrace, and one that reinforces regional innovation ecosystems and academic missions around impact.

However, this could also be symptomatic of a deeper structural gap: the absence of accessible, commercially-oriented shared infrastructure at the scale the sector requires^[7]. Academic institutions are configured for research and training, not commercial translation. Routing firms through them by default, rather than by design, can leave firms working with timelines and incentives that don’t match what getting a technology to market actually requires.

Finding 3: Biotech firms carry a higher-risk profile than deep tech peers

Biotech firms in the Horizon cohort show a somewhat different risk and innovation profile to their deep tech peers. They pattern with hard tech on IP intensity: 37% hold patents, against 34% in hard tech and just 23% in digital. Yet they draw 17% of their funding through Pillar III, the EIC’s higher-risk, breakthrough-oriented instruments, compared to around 5% for both digital and hard tech. This is consistent with a sector operating at earlier stages of the innovation cycle and with higher uncertainty than comparable cohorts.

Two years after a project starts, roughly one in four firms across all three sectors is loss-making, with rates of 24% in biotech, digital, and hard tech alike. That consistency suggests loss-making at this point is a feature of the Horizon deep tech cohort broadly, not a biotech-specific concern. The two-year snapshot cannot show how that loss-making picture diverges over time. 

Biotech’s patent intensity and its higher share of funding through EIC high-risk instruments point to a sector operating at an early, high-risk stage, the profile Horizon’s SME-oriented instruments are designed to support. Yet SMEs capture a smaller share of overall biotech funding than their headcount share or risk profile would suggest, and a smaller share than SMEs capture in either digital or hard tech. This raises a structural question about whether current Biotech programme design inadvertently favours larger, more established participants.

Finding 4: Biotech SMEs capture less funding compared to SMEs in digital and hard tech

SMEs make up 85% of biotech firms in the Horizon cohort, slightly higher than digital (83%) and hard tech (78%). But biotech SMEs capture only 22% of biotech funding, compared to 52% in digital and 35% in hard tech. Put the other way around, the small minority of non-SME biotech firms in the Horizon cohort capture nearly four euros in five. Some of this gap can be explained structurally: biotech Horizon funding tends to flow toward larger consortia and capital-intensive scale-up projects.

Finding 5: Horizon grants translate into revenue growth on average, but the picture diverges by domain

The descriptive profiles of the three deep tech domains differ, in firm characteristics, grant distributions, and position in the technology readiness cycle. But once a firm receives a grant, do these differences translate into differing financial outcomes for the firms?

We use a dataset built by researchers at Bocconi University^[8], linking Horizon Europe grant records to firms’ Orbis financial accounts. We restrict to firms with exactly one grant, so any revenue change can be traced to it, and compare revenue before and after, controlling for firm fixed effects and year fixed effects. What remains is our estimate of the grant’s effect.

Across the full sample, post-grant revenue is significantly higher than pre-grant revenue for firms active in deep tech Horizon projects. At the aggregate level, the grant does what the policy was designed to do. The sector breakdown diverges considerably across the three domains.

The post-grant effect for digital tech is significantly positive and persists across the full post-treatment window. This fits a sector with short paths to market, software-scale cost structures, and revenue cycles that sit comfortably within the grant horizon. The grant bridges a short financing gap to commercial viability, as the policy intends.

Hard tech is more selective. In aggregate the coefficient is negative and not statistically different from zero, but pulling subdomains apart we can see that advanced manufacturing, space & aerospace and MedTech & devices show significant positive effects, while the remaining hard tech sub-domains show no measurable revenue response. One plausible explanation is proximity to a downstream customer base. Where that customer base is identifiable and proximate, such as defence ministries, aerospace primes, healthcare providers, and established industrial supply chains, the grant translates into measurable revenue growth in the years that follow. Where the customer base is more diffuse, or dependent on long public-procurement timelines that operate on horizons much longer than the grant itself, the funding lands but it is not followed by revenue growth for the firm.

The biotech coefficient looks similar to digital, positive and significant. Stacking every firm’s trajectory on top of each other shows a more detailed picture: the post-grant effect is positive initially for biotech, broadly in line with digital, but it declines faster and turns negative by years four to five after grant receipt. This is consistent with what we know about biotech development timelines. Biotech firms operate on horizons long enough that, for a material share of grant recipients, the next technical, clinical, or regulatory milestone falls outside the window in which the grant is still visible in the firm’s accounts. When that milestone is missed, the follow-on private capital that would sustain revenue growth does not materialise, and the early effect of the grant erodes.

Finding 6: Independent SMEs gain more from grants than SMEs in corporate groups

We split firms into two groups using company ownership data from our financial database (Orbis): independent firms, which answer to no parent company, and corporate-group firms, which are subsidiaries or parent companies within a wider group. To see if there is a difference, we zoom into SMEs across both groups, so that firm size does not distort the comparison. A large company for whom a grant is a small fraction of its turnover would not be expected to show much of an effect.

Across deep tech, independent SMEs see their revenue rise by approximately 32.7% after receiving a grant. SMEs owned by a corporate group also see a rise, but a smaller one, at 18.4%.

Among biotech SMEs, independent firms see a 24.7% increase. Biotech SMEs owned by a corporate group show no significant change in revenue after receiving a grant.

These findings mirror those of the Institute for European Policymaking at Bocconi University (IEP@BU), whose analysis of all Horizon grants found the same pattern: EU grants boost revenue more for independent SMEs than for SMEs owned by corporate groups^[9]. IEP@BU proposes an explanation based on financial need: subsidiaries are unlikely to be financially constrained, since a parent company can plug gaps that a grant would otherwise fill. For a standalone startup, the same grant can determine whether the firm and its ideas survive at all.

Corporate-group biotech SMEs still take home the larger share of biotech SME funding, 66% against 34% for independent biotech SMEs, despite the evidence that those grants don’t measurably translate into revenue growth. This skew towards corporate-group recipients holds across deep tech more broadly: corporate-group SMEs receive 64% of SME funding in digital tech and 73% in hard tech. A shift of even a modest share towards independent SMEs could plausibly generate more output for the same level of public spending.

What this means for policymakers

Instrument calibration is key for biotech

To illustrate our point, the EIC accelerator has a €10 million equity ceiling and a €2.5 million grant ceiling, which together cap the public capital any single firm can receive. This may be adequate for a digital firm at median revenue with a short path to market, but the same ceiling often falls short for a biotech firm at the same revenue level, facing multi-year regulatory approval, significant pre-commercial capital expenditure, and no near-term return. The instrument is calibrated to a generic deep tech average, and applying that average uniformly, rather than to each sector’s development logic causes a miscalibration.

Recent additions partially acknowledge the gap. The Strategic Technologies for Europe Platform (STEP) Scale-Up, launched in 2025, channels equity-only investments of €10-30 million through the EIC Fund into scale-ups across digital, clean tech, and biotech, aiming to catalyse total funding rounds of €50-150 million. Advanced Innovation Challenges, an ARPA-style pilot launched in the EIC’s 2026 work programme, tests a different design: challenge-driven, demand-led funding staged across two phases, targeted at fields where Europe is strong in research but slow to commercialise. Both instruments offer larger or differently structured funding than the EIC Accelerator’s fixed grant-and-equity ceilings, consistent with an emerging view that grant-based support alone may not carry firms across the commercialisation valley. Whether their calibration reflects biotech’s specific cost structure and development timelines, or still leans on a broader deep tech average that systematically underestimates what the sector requires, is something only time and further data will show.

Programme design features compound this misalignment, such as the EIC one-grant rule, which limits firms to a single Accelerator award over the programme period and is consistent with a model in which firms progress linearly from late-stage development to commercialisation. This assumption is more likely to hold in digital than in biotech where firms at TRL 6–8 may still be several years and millions away from regulatory approval or revenue generation; a single intervention may not bridge the financing gap. The structural risk is that firms exit public support before private capital is willing to enter, and in biotech that private capital is scarce to begin with: US biotech venture capital investment has run consistently ahead of the EU’s over the past decade, reaching up to ten times the European level at its 2021 peak^[10].

The same instrument interacts differently with firms depending on development model, capital intensity, and time to market. In biotech, where early-stage research, high-risk innovation, and capital-intensive scaling are tightly interwoven, these differences become particularly pronounced.

Policy is beginning to reflect this. The proposed €10 billion investment pilot under the EU Biotech Act, implemented with the EIB Group, explicitly aims to tailor financing to biotech-specific risk profiles rather than the generic deep tech average. It is likely to be particularly relevant for industrial and non-health biotech, where firms face the same capital-intensive development requirements as health biotech but without the regulatory certainty or specialist investor base.

The broader implication is that funding effectiveness depends as much on the volume of support as on alignment with sector-specific development dynamics. Where these dynamics diverge, as they do in biotech, a more differentiated approach to instrument design is required.

Funding should follow a strategic and compounding architecture

Biotech does not move through neat stages from research to engineering to market, rather scientific challenges run through the entire development cycle. Scaling a biological system, optimising fermentation, or qualifying an engineered organism each requires solving real scientific problems, not just engineering up what worked in the lab^[11]. Funding architectures that assume otherwise are likely to impede technology translation^[12].

So, if biotech does not develop linearly, neither should funding. A European perspective has to support the whole bio-stack in parallel. For industrial biotech that looks like: the strategic question of which feedstocks European biological systems can viably grow on, the engineered organisms themselves, the predictive digital systems that anticipate how those organisms will behave at industrial scale, the fermentation and processing steps that scale them up, and the analytical methods that verify what they produce. Funding one layer while starving another does not move the stack forward, it just shifts where the bottleneck sits (read more in our Call For Evidence for Biotech Act II).

Coordinating this kind of parallel progress requires a standing body with the mandate and continuity to track and update all five layers together, rather than ad hoc oversight of each in isolation. The US biotech roadmap points to Sematech, the long-running US semiconductor industry consortium, and the UK Synthetic Biology Leadership Council, established in 2012 to refresh the UK roadmap on a rolling basis, as two working precedents for this kind of body^[13]. This is the kind of deliberate and continuous alignment of EU policy, investment, and innovation we called for in Towards a European BioPower.

For European policymakers, the case for bio-stack thinking maps onto two priorities already in motion: Strategic autonomy in biotech depends on coordinated capabilities across the stack, not isolated strengths in individual layers. A country that funds organism design without fermentation infrastructure, or fermentation without analytical methods, must still depend on and import the missing pieces. And planning public funding across the bio-stack prevents waste: investment in any single layer pays back only if the surrounding layers can absorb it. This coordination is not an argument for trimming Pillar I or compromising European Research Council independence. Investigator-driven frontier research is the foundation the whole stack depends on. Differentiated instruments elsewhere in the stack exist to capture and support that value, to help realise it, not to draw funding away from Pillar I.

The EIC suite operates on a different logic. The Commission reviews priorities and themes annually through Challenge calls and work programmes – but to what extent does this include strategic full bio-stack thinking? Horizon funding architecture has continued to evolve: STEP Scale-Up arrived in 2025, Advanced Innovation Challenges piloted in 2026. But the underlying parameters of the EIC instruments and STEP Scale-Up (TRL bands, ceilings, eligibility rules, the one-grant constraint) sit in the multi-annual programme, with no standing body reviewing whether they keep pace with the development logic of the quickly advancing technologies they are meant to support. A standing mechanism to ensure this learning loop is implemented would increase the effectiveness of EU public funding.

The Biotech Act creates an opening to change this. Its first, health-focused proposal (COM(2025) 1022, December 2025) introduces “health biotechnology strategic projects” and “high impact health biotechnology strategic projects”: designations that recognise priority projects and unlock fast-track permitting, coordinated support, and easier access to EU funding. The high-impact tier is reserved for projects that demonstrate, “by virtue of [their] scale, scope or cross-border relevance, a strong systemic and catalytic potential within the Union’s biotech ecosystem to accelerate innovation and enhance the translation of research into market applications”^[14]. This systemic, ecosystem-level framing is the logic the bio-stack requires. Four features will determine whether it delivers on that logic:

1. The designations cannot remain confined to health. As currently drafted they reach only health biotech. The forthcoming non-health package (Biotech Act II) expected at the end of 2026 should equally integrate systems-level and bio-stack logic.
2. The designations need continuous updating, so the system tracks a fast-moving technology base rather than locking in yesterday’s priorities.
3. They need to span the full technical stack, not name isolated projects. For example, for non-health biotech, that would mean feedstock science, fermentation, processing, organism design, and analytical infrastructure advancing in parallel.
4. There needs to be a robust feedback mechanism into instrument design itself, so the EIC and new instruments match the development logic of the technologies they support, rather than the deep tech average that this analysis has shown does not fit biotech.

Biotech Act II should borrow ambition and learnings from the Chips Acts

In 2020 and 2021, the global chip shortage turned a long-standing strategic weakness into a visible crisis. The US made 37% of the world’s chips in 1990; by 2020 its share had fallen to 12%^[15], and the EU’s share of global manufacturing capacity stood at around 9%^[16]. Car plants sat idle, and in Germany the hit to car production alone is estimated to have cut GDP by around 1.5% in the first nine months of 2021^[17]. Both jurisdictions reached the same conclusion: a foundational industry had moved offshore, and reversing that required public money on a bold, industrial scale.

The CHIPS and Science Act passed in the US in August 2022, carrying $52.7 billion. The EU Chips Act followed in September 2023, mobilising up to €43 billion^[18]. Both take the same shared architecture approach: a significant portion of the money targets capability the whole sector can draw on, rather than individual companies. This includes pilot lines that bridge laboratory process development and industrial-scale fabrication, the data that accumulates across them, and the workforce trained to operate them. The EU launched four pilot lines, each with an explicit mandate to serve start-ups and SMEs. The research was being done in Europe, the manufacturing had gone elsewhere, and bridging the two was too expensive and too risky for any one firm to attempt alone.

Is there evidence this works? The Chips Act’s record separates cleanly into two halves. The European Court of Auditors found progress clearest under the pillar housing the pilot lines and shared research infrastructure, while the firm-level manufacturing-subsidy pillar saw slow uptake. The same audit concluded the Act is very unlikely to reach its 20% market-share target by 2030, a target the subsidy machinery was meant to deliver. When the Commission proposed Chips Act 2.0 in June 2026, it kept and expanded the pilot lines and overhauled the target-and-subsidy design. The half that worked is the half biotech should borrow^[19].

This same logic applies to biotech – we propose a three pillars structure where the EU should concentrate public investment to build this shared-capability layer: physical infrastructure, intelligent digital infrastructure, and knowledge infrastructure.

1. Physical infrastructure – the foundational backbone for scale-up support

A shared European scale-up infrastructure would let any SME with a promising fermentation process or a novel production host work through the engineering and scientific challenges that emerge at each scale transition, generate the data needed to attract follow-on capital, and leave with a validated route to market. This is what shared, affordable biomanufacturing infrastructure would offer: access to facilities a small firm could never build on its own.

It would also de-risk the sector for private investors. A process proven at pilot scale, with the performance data to show it, is a far more fundable proposition than a promising result in a lab flask, which is often where the capital hesitates. This breaks a funding deadlock: the proof investors want is expensive to produce, and a shared facility does not make it free, but by spreading the cost across many users it brings that first proof within reach of firms that cannot yet raise against it. A shared site offers more than equipment – firms tap into the operators, process engineers, and hard-won troubleshooting know-how already there.

This is most urgent and impactful for the newest biology, where nobody knows how a novel process will behave outside the lab. Take AI-designed biologics: proteins and other molecules generated by algorithms rather than found in nature.^[20] However well the design performs on screen, no one knows how the process will behave in a fermentation tank. Or engineered chassis cells, microbes whose DNA has been rewritten to turn them into tiny factories, which tend to lose those modifications over a long production run: the cells that abandon the costly engineered traits grow faster and take over.^[21] This is called selection pressure, evolution at work in real time! Or new manufacturing paradigms, such as continuous bioprocessing, which keeps a bioreactor running for weeks instead of restarting after each batch, but few processes have made the transition from laboratory demonstration to commercial scale.^[22] Other approaches move away from the standard microbial workhorses like bacteria and yeast fed with sugar. Cell-free systems skip living cells entirely, using only the molecular machinery extracted from them^[23]; and gas-fermenting microbes are being engineered to grow on industrial CO₂ or waste gases rather than sugar, turning emissions into chemicals and fuels^[24]. Both are too new at industrial scale for anyone to have worked out how to run them reliably. In every case above, the science and engineering sit at the frontier and the manufacturing knowledge needed to run these processes at scale is still in development. This makes shared learning particularly high-impact as compared to more established processes.

Shared not-for-profit (NFP) infrastructure is the most direct route. Bio Base Europe Pilot Plant in Ghent has run as an open-access, NFP facility for SMEs since 2008. It has completed more than 900 confidential bilateral projects with over 300 companies and research institutions worldwide, and has experience with over 90 consortium-based projects, including more than 30 BBI JU and CBE JU projects. That track record makes shared facilities a proven mechanism for catalysing EU funding and de-risking innovation at demonstration scale, not a marginal add-on. Its infrastructure is modular: individual process steps, such as pretreatment, fermentation, and purification, can be run at different scales and recombined, rather than the whole line being fixed to one size. This attracts international firms to Europe for scale-up. Cemvita, a US firm, ran its fermentation process through 2-litre, 30-litre, 1,500-litre, and 15,000-litre stages before reaching a 75,000-litre demonstration run, climbing the entire scale-up ladder at Bio Base Europe without switching sites or partners.^[25] Europe already has an asset that works, is competitive, and attracts international players. The problem is that it is one of a handful of legacy facilities, loosely networked through initiatives like Pilots4U, rather than a funded, expanding public system.

These kinds of public SME-facing infrastructures are fragile. The UK’s Vaccine Manufacturing and Innovation Centre was funded with £250 million by 2021, then sold off before completion in 2022 when fiscal priorities changed, a decade of strategic intent undone by a single change of government.^[26] They are fragile operationally too: equipment ages faster than budgets can replace, skilled operators leave for better-paid industry jobs, and pressure to recover costs pushes access toward the established firms that can afford it, rather than the early-stage firms the facilities were built to serve.

On the private side, only a handful of Europe’s 20 to 30 industrial biotech Contract and Manufacturing Organizations (CMOs) are equipped for large-scale fermentation, and those that are tend to prioritise established, long-term clients, leaving the most innovative early-stage players stranded.^[27] Access also depends on process fit: a startup’s downstream process must be compatible with the CMO’s existing infrastructure, no CMO will build a new downstream processing line to accommodate a startup, however promising its future business looks. This is the gap pilot plants like Bio Base Europe are designed to close. As open-access, mission-driven facilities, they support unproven and non-standard processes that a commercially-driven CMO cannot. But no single facility can offer every process fit. Europe needs complementary specialisations across facilities, covering different feedstocks, fermentation types, and downstream processes, so a start-up can find a technical match somewhere in the system. Biotech Act II should build this with three lessons in mind: long-term political and financial commitment so facilities are not sold off when budgets tighten, governance that keeps access open to early-stage firms rather than prioritising larger repeat customers, and capacity that keeps pace with the frontier.

2. Intelligent digital infrastructure – data as the new shared competency

A pilot plant’s output is not only a validated process, every scale-up campaign also produces something else: operational data that no single SME can produce alone, and that no individual firm has incentive to share. Run across a network of facilities, that data becomes a valuable shared record of what works, what fails, and under what conditions, accumulating across users. This is federated bioprocess intelligence, and it improves as more campaigns run through the network. Today the opposite happens: each firm keeps its results in its own format, and the same costly failures are rediscovered across the sector, one company at a time.

In Towards a European BioPower, we make the case for shared research infrastructure across member states, from cloud labs to biofoundries, as the way to “transform Europe’s fragmented biotech landscape into a cohesive ecosystem”. That argument is about accelerating the Design-Build-Test-Learn cycle at lab scale, biotech’s core engineering loop: design an organism or process, build it, test how it performs, and learn from the result before the next iteration.^[28] This logic can be carried downstream, to the pilot-scale manufacturing infrastructure where the loop runs again, this time proving the process at industrial volume rather than at the bench. Because scale-up processes are run on common infrastructure under common protocols, the learnings are transferable through shared data standards rather than trapped in incompatible in-house datasets, and a single European record accumulates in place of thousands that are not interoperable.

This matters most for the frontier processes described above, which do not have decades of industrial experience to fall back on. For these, the shared data learnings are the manufacturing roadmap. Building the capacity to capture and pool it is work no single firm will undertake, because no single firm captures the return: precisely what public investment exists to provide.

This thinking is already being implemented in the semiconductor sector: In the US, the National Semiconductor Technology Center pairs a shared design and data layer with common prototyping and production lines, so knowledge on research through to scaleup processes accumulates centrally rather than dissipating.^[29] Biotech Act II could set the political commitment to build the equivalent for biomanufacturing.

3. Knowledge infrastructure – talent at home and for export

Shared pilot plants and the data they generate are part of successful biotech scale-up. The other part is the knowledge infrastructure that runs them: the trained operators, engineers, and scientists who resolve the scientific and technical problems that arise when laboratory processes meet industrial scale.

Ireland’s National Institute for Bioprocessing Research and Training, established in 2011 with €57m of IDA Ireland funding, is the clearest working example. NIBRT is a biopharma institute running as a four-university consortium (UCD, Trinity, DCU, IT Sligo). The structure plays to biotech’s already-deep university dependency (50% median consortium share in our data, the highest of any deep tech sector) and orients it toward industrial translation. Its facility is a working bioprocessing plant replica: single-use bioreactors, automation stacks, QA workflows identical to industry. Critically, NIBRT is not only a training centre, it also functions as an applied R&D test bed, so cohorts do not only learn to operate established processes but learn to develop new ones. It trains over 4,000 operators and engineers per year.

Ireland is the world’s third-largest pharmaceutical exporter, with €116bn in annual export revenue and 85,000 sector jobs. NIBRT is consistently named as the reason multinationals chose Ireland over comparable jurisdictions: the structural answer to where a manufacturing operation will source its workforce.

NIBRT’s curriculum, facility design, and certification regime have been licensed to the Jefferson Institute for Bioprocessing in Philadelphia ($10m partnership) and to K-NIBRT in Songdo (2,000 Korean trainees per year). Each deal generates revenue, extends a global professional network anchored in Ireland, and embeds NIBRT-pioneered processes as the operating standard in the receiving jurisdiction.

Currently, no equivalent exists for European industrial biomanufacturing. Training in these domains happens ad hoc, inside individual firms, on equipment configurations that do not transfer. Whichever jurisdiction builds this institutional model first will set the standards the rest of the sector operates within.

Europe is structurally well-positioned to be that jurisdiction. The bloc hosts the densest concentration of research-intensive universities in the world, and biotech already runs through them more deeply than any comparable deep tech sector. The four-institution consortium model that delivered NIBRT is replicable at European scale several times over, with regional specialisation across Member States. The talent base is in place, what is missing is the institutional design that converts academic capacity into industrial capability.

The infrastructure pillars compound to bring returns for public spending

The three pillars deliver more together than separately. A trained workforce, physical scale-up infrastructure, and accumulated process intelligence in one place are what a firm needs to generate the data that convinces private investors a project is de-risked, and what the current European ecosystem fails to deliver under any existing instrument.

The leverage on Horizon spend follows from this. Each grant today funds one project at a time, with the learning contained inside its firm or consortium. A federated pilot plant infrastructure changes that. Every campaign that runs through it, whether Horizon-funded, privately backed, or nationally supported, deposits knowledge, protocols, and more experienced operators back into the platform for the next user. For a sector struggling at the science-to-commercial transition, this determines whether public investment compounds or dissipates.

Conclusions

Findings: Horizon biotech firms resemble digital tech firms in size and revenue, but their development process runs on hard tech’s timelines and capital needs. Biotech firms skew to earlier-stage, more research-intensive work than either peer sector, with heavier university dependence, fundamental science funding and higher-risk EIC funding. Grants raise revenue on average, but the effect fades within four to five years, faster than digital tech, tracking biotech’s longer path to the next milestone. Independent biotech SMEs see revenue rise 24.7% after a grant; SMEs owned by corporate groups show no significant change. Corporate-group SMEs receive 66% of biotech SME funding against 34% for independent SMEs, and non-SME firms take close to four euros in five of all biotech funding. Funding concentrates in the segments showing the weakest measurable return.

Conclusion: Horizon’s instruments seem to be calibrated to a generic deep tech average. Our data points towards a mismatch between that average and biotech’s cost structure and timelines.

Suggestions: Calibrate instruments to biotech’s actual development logic rather than the deep tech average, with a mechanism to update that calibration as the field evolves and accelerates. Build bio-stack logic deep into EU funding logic, so instrument design funds R&D across the whole bio-stack, working toward solving (particularly for industrial biotech) scale-up challenges and increasing cost competitiveness. Bio-stack logic recognises that science-to-market challenges run throughout biotech development and that its layers are interdependent, a concept we put forward in our Biotech Act II Call for Evidence Submission. Build shared, EU-funded infrastructure across three pillars: physical scale-up facilities, a federated data layer, and training institutions modelled on NIBRT. This lets public investment compound across users instead of dissipating project by project. Direct a larger share of investment toward independent SMEs, the segment where grants demonstrably convert into revenue growth.