大学発ベンチャー、システムを変革を

“An organisation that hires people to do something stupid will tend to continue to do it, because the people they’ve hired will lobby for it in order to keep their jobs.”— Paul Graham (@paulg), January 2023I retweeted this in 2023. I’m thinking about it again now, because nowhere does it ring truer than in the UK’s university tech transfer and biotech accelerator ecosystem.There is a myth at the heart of UK science policy. It goes like this: Britain has world-class research, we punch above our weight, we just need to get better at commercialisation. The first part is true. The second is not. And conflating the two has let a failing system off the hook for decades.The UK is genuinely exceptional at one thing: publishing papers. First in the world for quality of academic publications by field-weighted citation impact. Third globally for total share of publications and citations, behind only the US and China. Producing 57% more academic publications per capita than the US. The Research Excellence Framework rated 84% of university research submissions as world-leading or internationally excellent.[1]Publishing papers is not building companies. Citation counts are not revenues. And the system that was supposed to bridge that gap, technology transfer offices, accelerators, spinout infrastructure, has instead become very good at generating its own metrics while the underlying problem persists.The UK ranks sixth out of G7 nations for employment in medium and high technology manufacturing. Technologies invented here are routinely commercialised and manufactured elsewhere. The UK was a leader in drug discovery and lacked the capacity to scale vaccine production during a pandemic. This is not a punching-above-your-weight story. That is a structural failure with a long paper trail.[2]The problems start at the point of formation. Oxford was, until recently, taking up to 50% equity in spinouts before any investment was made. After a 2021 reform, it settled on a “take it or leave it” flat 20%, still described by founders who’ve been through the process as one of the worst deals available. One Oxford quantum spinout made headlines when founders were left with less than 5% equity between them after their Series A. Oxford’s TTO was publicly attacked by VCs on X as recently as 2024, who argued its commercialisation terms were directly responsible for spinouts failing to attract funding. And Oxford is supposed to be one of the better ones.[3]The 2025 Spotlight on Spinouts report, the most authoritative annual survey of UK spinout activity, found that just ten universities are responsible for 53% of all UK spinouts. The average university equity stake fell to 16% in 2024, down from 22% the previous year, in response to sustained pressure from founders and investors. Progress, yes. But the fact that halving equity stakes is considered a major achievement tells you where the baseline was set.[20]In numbers:16% average UK university equity stake in spinouts in 2024. a decade low, down from 22% in 2023[20]67% of UK spinout deals take more than six months to close[4]30% of founders waited a year or more to strike a spinout deal with their university[5]53% of all UK spinouts produced by just ten universities[20]The University of Southampton only lowered its standard one-third equity stake in May 2024, after a decade of taking it.[7] These are not outliers. They are the norm across the majority of UK universities, most of which lack the commercial expertise or founder-facing culture to behave any differently.Public funding for early-stage founders exists. Innovate UK, Biomedical Catalyst, and others, but it is fragmented, oversubscribed, and poorly targeted. And even when it works, it predominantly de-risks the science rather than building the commercial capability founders actually need. The UK’s proof-of-concept funding. The critical bridge between academic research and an investable company amounts to £40 million over five years across the entire country.[8] For context, five universities in Belgium’s Flanders region receive more than that every single year from the regional government alone. Grant funding is not absent, but it is neither sufficient nor well-connected to what founders building real businesses require.None of this comes cheap. The Higher Education Innovation Fund. The primary public mechanism through which TTO operations are sustained in England distributed £260 million in 2023–24 alone. Across the four UK nations, the total is higher still, and this sits alongside separate research council funding, Innovate UK grants, and the internal university budgets that subsidise TTO staff costs where HEIF falls short. Cambridge’s TTO alone costs approximately £2.5 million per year to run. Multiply that across 130-odd institutions. many with far lower commercial returns. and the taxpayer is funding an enormous infrastructure whose median output, by the sector’s own data, barely breaks even.[19]“If you aim to show something, you invariably do” is something I originally coined to reflect the challenges when bringing the engineering mindset into science, but it seems appropriate here. TTOs are asked to show spinout numbers, so they produce spinout numbers. They are asked to show licensing income, so they protect IP that maximises licensing income. They are not asked to show whether founders succeeded, whether technologies reached patients, or whether the public got a return on its investment. So they do not show those things, and over time, they stop thinking about them.TTOs have been structured to maximise revenues rather than maximise the number of technologies that actually reach the world. The result is a system optimised for protecting IP and extracting licensing fees, not for getting things built. They became gatekeepers, not facilitators.[9] Research across 98 structured interviews at US and UK universities found the same pattern: bureaucratic inflexibility, poorly designed reward structures, and ineffective management were the defining features.[10]The government’s 2023 independent review acknowledged the problem.[11] Some universities have moved: Southampton cut its equity stake, Imperial introduced its Founders Choice programme. But voluntary improvement is slow when the incentive to change is weak, and the people running the current system have jobs to protect.Here’s the data point that should make every TTO director and accelerator manager uncomfortable: the ecosystem they claim to serve is largely building itself without them.The Royal Academy of Engineering’s State of UK Deep Tech report, the most comprehensive analysis of the sector, found that university spinouts account for just 34% of the UK’s deep tech ecosystem. The remaining 66% are independent startups: companies founded by people who didn’t go through a TTO, didn’t wait for a licence, didn’t sit through a negotiation about equity. They just built something.[21]This is the number that exposes the fundamental mismatch in UK innovation policy. The vast majority of public debate, government review, and institutional energy is directed at the minority route. Spinouts get the reviews, the policy papers, the equity debates, the HEIF funding, and the TTO headcount. Independent founders, who make up the majority of the deep tech ecosystem, get comparatively little. We have built a system to optimise the 34% while largely ignoring the 66%.A systematic analysis of biotech companies that reached IPO reinforces the point: only 3% of CEOs came from academia. The overwhelming majority came from industry, pharma (39%), biotech (29%), or financial services (27%). The founders had PhDs, yes. But they built their companies from industry experience, networks, and capital, not from a university tech transfer process.[15]In the UK, the pattern holds. The companies most often cited as evidence of UK deep tech strength, including DeepMind’s founders, Wayve and Synthesia, emerged from university education, not university commercialisation infrastructure. They were startups, not spinouts. They owned their IP. They didn’t wait six months for a TTO to process their deal. The system takes credit for the talent it has educated. It had little to do with what those founders actually built.The deeper problem is cultural. The policy conversation treats the spinout as the canonical form of deep tech company creation, and the TTO as the canonical gatekeeper of innovation. But innovation does not come from organisations. It comes from people. Specific people with specific insights, at specific moments, willing to take a specific risk. The question that should anchor every conversation about the UK’s deep tech future is not “how do we create more spinouts?” but “how do we find the people building things and get out of their way?”Meanwhile, some universities have been actively gaming the spinout numbers. Research in the US found that in 2008, one of the universities reporting the most spinout formations had the majority of those companies listing the TTO’s own address and phone number as corporate headquarters, with the TTO director listed as corporate officer. No real business. Just entries on a spreadsheet that justified the TTO’s budget for another year.[16]The UK government’s own data shows there are 1,337 active spinouts across the country, with 204 exits since 2014. That sounds impressive until you note how concentrated those exits are among a handful of institutions, and that the definition of “active” does not necessarily mean functioning, funded, or building anything. It means the company hasn’t been struck off.[17]The founders building genuinely innovative UK biotech companies, in cell therapy, synthetic biology, diagnostics and biomanufacturing, are increasingly doing it as independents. They’re raising from angels, syndicates, and the occasional enlightened VC. They’re not waiting for a TTO to license their IP or an accelerator to run them through a pitch deck workshop. The system wasn’t built for them. It was built for the people running it.Accelerators are, in many cases, the consumer-facing version of the same problem.The UK life science accelerator landscape has actually “grown”: 63 offerings in 2024 versus 62 in 2021. That sounds like progress. But in the same period, 22 offerings were retired and 23 new ones launched. Considerable churn, not consolidation.[12] Many exist because they have a grant to spend. When the grant runs out, they close. A new programme launches with a new grant. The staff moves across. The cohort fills up. The press release goes out.The pitch is compelling: mentorship, lab access, warm introductions to investors. The reality is often workshops on pitch decks, demo days that go nowhere, and equity taken from founders who are still two years from having anything to show investors.Innovate UK, the UK’s primary mechanism for early-stage innovation funding, released only £200 million in innovation grants in the 2024 fiscal year, against an expected £1 billion. A five-times shortfall. The government shutdown across Q2/Q3 2024 is blamed. Meanwhile, the accelerators continued running their cohorts.[13]What accelerators are very good at is producing the appearance of activity. Application portals. Selection panels. Cohort announcements. Graduation ceremonies. Press releases. Each of these requires staff, and that staff needs to justify its existence, which means running the next cohort.The people hired to run accelerators will advocate for accelerators. They’ll find evidence that they work. They’ll point to their one or two portfolio companies that raised a Series A. They won’t mention the thirty that quietly dissolved.Biotech is not software. You cannot iterate to product-market fit on a £50k grant. A cell therapy platform needs GMP manufacturing. A diagnostics company needs clinical validation. A bioproduction tool needs real customers with real workflows, not a demo day audience.The UK’s venture capital landscape is already heavily concentrated geographically, with early-stage biotech founders outside a handful of cities finding the market structurally hostile. No UK biotech went public in 2025, the third consecutive year without a listing. Venture capital investment fell 13.2% year-on-year in 2025, reaching £1.79 billion across just 58 deals. Funding was increasingly concentrated into a small number of large rounds, with no recovery in mid-sized scale-up financings. Early-stage founders, the people accelerators claim to serve, are precisely those being left behind.[14]Into this environment, UK biotech founders are also contending with TTOs that take too much equity too early, delay deals for months or years, and install their own people on boards. Then they’re pointed towards accelerators that run workshops. Then they’re told to apply for grants in a year when Innovate UK’s budget collapsed by 80%, despite Biomedical Catalyst and other routes remaining technically open.The mismatch between what the system offers and what founders actually need is vast. And it’s not closing, because the people running these programmes are not rewarded for closing it. They are rewarded for filling in the next cohort.Right now, the UK is spending around £300 million a year on salary costs and direct public funding for an ecosystem that has become a job creation system funded by the taxpayer, but seems to have forgotten that the whole purpose is not to create jobs, it is to build industries that help solve real problems and grow the economy at the same time.To put that £300 million in context: it is more than the entire UK proof-of-concept funding commitment over five years, spent annually, largely on a system that serves the 34% of deep tech companies that go through formal spinout routes. It is built on top of the £260 million per year distributed through HEIF, supplemented by separate research council funding, Innovate UK grants, and internal university budgets. Add salary costs across an estimated 2,000 to 2,500 TTO and knowledge exchange professionals and the staff running 63 life science accelerator programmes, and the total picture is of an enormous publicly funded infrastructure whose median output, by the sector’s own data, barely breaks even.Imagine for a moment what we could do if even a fraction of that £300 million went directly to people building things rather than the bureaucracy around them.This will not be a popular suggestion, but these departments and their budgets should be significantly reduced, and new bureaucracy-light, staff-light mechanisms built to fund founders directly. Compare that with ARIA. The Advanced Research and Invention Agency operates with just 53 staff, deploying a minimum of £1 billion over the 2025 to 2029 spending review period with a high-risk, high-reward mandate and minimal process overhead. That is a ratio of capital to headcount the TTO ecosystem cannot come close to matching. Sadly, I will also argue in a future piece that ARIA, too, has already begun to exhibit the same organisational dynamics as Innovate UK, the pull towards process, the accumulation of staff, the gradual drift from mission to maintenance. But the model is closer to right than anything the TTO world has produced.Expanded founder fellowships. They already exist, but few provide direct founder support and many fund the institutional overhead that already consumes too much resource. The Royal Academy of Engineering’s Enterprise Fellowships provide direct salary support to deep tech founders. They work precisely because they fund the person, not the organisation around them. They are also chronically small in scale. Expanding these programmes significantly, and creating equivalents assessed on commercial and technical milestones rather than publication outputs, would get money to founders faster and more efficiently than any accelerator cohort. With respect specifically to biotech, there is at least one program that appears from the outside tailored to founder scientists, and this is Wilbe’s program, where the focus is on helping the founder rather than the infrastructure around them.This aside, the question is not whether founder fellowships work. It is why we are not doing far more of them where founders are directly allocated funds. Imagine what we could do instead of funding around 50 such fellowships a year (when combined with institutional awards such as UKRI Future Leaders Fellowships), we could fund 500. Imagine the number of real-world, commercially focused entities that would emerge here in the UK.Milestone-based direct grants. Fund founders against commercial outcomes, not research outputs. Did you get a paying customer? Did you hit a manufacturing milestone? Did a hospital agree to trial your diagnostic? These are fundable events. A PowerPoint of research aligning with a Gantt chart is useful to show that work has been executed, but it is not a commercial outcome.Expanded SEIS and EIS with reduced friction. The Seed Enterprise Investment Scheme is one of the most effective founder-facing funding tools the UK already has, offering meaningful tax relief to early investors. Raising the limits, simplifying the process, and actively promoting it outside established investor networks would get more capital to more founders at a lower cost than any accelerator programme.Reformed co-investment. The UK already runs government-backed co-investment schemes that match public capital alongside private investors, and in principle, this is a sound model, it leverages private judgment rather than replacing it with process. The problem in practice is that these schemes are slow, bureaucratic on the public side, and often arrive after the moment of need has passed. Investors are also sometimes sceptical: if the government is matching their bet, they want to know whether it is because the deal is strong or because it qualified for a programme. Reformed co-investment would need to be faster, lighter on documentation, and structured so that the public capital genuinely follows the private decision rather than running alongside its own separate assessment process. Done well, it is one of the most capital-efficient mechanisms available. Done badly, it is another queue to join.Public procurement as a first customer. The NHS and other large public bodies could be far more deliberate early buyers of UK deep tech. A committed first customer is worth more to most founders than six months in an accelerator. The incentive and procurement structures to make this happen are largely absent, which is a policy choice, not an inevitability.The common thread across all of these is the same: align the incentive with the outcome, not the activity. Fund the founder, not the organisation claiming to support them. Measure success by whether the thing got built and whether someone paid for it, not by how many cohorts ran or how many patents were filed.Until that changes, most TTOs and accelerators will continue doing what they do. Not because it works. But because the people hired to run them will lobby to keep running them.References[1] Cambridge Industrial Innovation Policy. Does the UK’s scientific research translate into industrial success? UK Innovation Report 2024. ciip.group.cam.ac.uk[2] Cambridge Industrial Innovation Policy. Does the UK’s scientific research translate into industrial success? UK Innovation Report 2024. ciip.group.cam.ac.uk[3] Sifted. How bad are Oxford University’s spinout policies? June 2023. sifted.eu[4] Odelle Technology. The Role of Technology Transfer Offices in Innovation and the Challenges They Face. July 2024. odelletechnology.com[5] Onward. Venturing Out: Science Superpower. June 2024. ukonward.com[6] Tech Funding News. University spinouts: Should universities take lower equity stakes? August 2025. techfundingnews.com[7] Global Venturing. UK universities lower spinout equity stakes amid calls for growth. July 2024. globalventuring.com[8] Global Venturing. Make sure U

※ 著作権に配慮し、引用は冒頭3段落までです。続きは元記事をご覧ください。