The open database for university spinouts
Help us level the playing field by shedding light on how entrepreneurial researchers at universities get to spin out their inventions.
The spinout opportunity, the deal-making problem…and a policy solution
Published by Nathan Benaich and Othmane Sebbouh on 30 June 2022.
In an effort to make spinout deal terms more transparent, we have released the spinout.fyi database, along with an analysis of the data. The spinout.fyi database will always be open and free. In its first version, the database contains 143 records from 71 institutions around the world. We encourage all spinout founders to anonymously share their experience and spinout deal terms at https://www.spinout.fyi/#survey.
University spinouts represent a mere 0.03% of the UK’s overall universe of companies (4.7 million). Normalising against the number of academics in the UK, that’s one spinout per 195 academics. By contrast, the UK has one company per 10 working-age UK citizens (16-64 years old). One could therefore say that academics are ~20 times less likely to form a spinout than your next door neighbour. Given the world-class research output from the UK’s celebrated academic institutions and the nation’s desire to attain tech sovereignty, shouldn’t we be doing far better than this?
The UK’s spinout opportunity
California-based Databricks is today the $38 billion exemplar of the modern software company success story that was unthinkable a few decades ago. Starting off as an open source project, Databricks quickly made its way to partnering with some of the largest cloud computing providers including Microsoft and Google, and serving more than 5,000 customers.
But Databricks is an unusual company in another way too. The business is a spinout from UC Berkeley’s now illustrious RISELab (formerly known as the AMPLab). This university lab has created some of the most used tools in distributed computing like Apache Spark and Mesos. Given Databricks’ valuation, the university’s equity share has already rewarded it handsomely. But there is a more surprising side to this success story: UC Berkeley announced in June 2021 that it had received three gifts of $25M each, two of which came from Databricks co-founders Scott Shenker and Ion Stoica.
For a long time, this story would have been more likely to unfold in the US than in the UK. But the UK’s startup scene has exploded in the last decade and is increasingly populated by companies that bring scientific inventions to real-world industries, be it artificial intelligence (AI), biotechnology or quantum computing. Many of these companies are founded by entrepreneurial researchers at universities who have the ambition and energy to lead their inventions beyond publications and into the real-world. A notable example is Exscientia, the recently IPOed biotech company using AI for drug discovery, which spun out of University of Dundee in 2012.
Why do spinouts matter to our technology ecosystem?
We are currently in a global race for technology leadership and sovereignty. Spinouts represent a key vehicle to deliver on this mission. AI, for example, is considered by many to be the transformational technology of our time. As a scientifically-rooted field, many companies that establish leadership with AI-first products emerge from university ecosystems. Indeed, according to Beauhurst, 4.3% of UK AI companies are formed by university spinouts. However, spinouts account for only 0.03% of all UK companies. More generally, spinouts are the most direct and efficient mechanism to translate scientific discoveries into tangible improvements for everyday life. As mentioned in Sahaj Sharda’s essay “Seize the University Patents”, spinouts are the ultimate justification for publicly funded research.
Spinouts and talent are two essential parts of a positive feedback cycle that will allow us to achieve technological leadership. Universities that enable the successful creation of spinouts will attract academic talent with entrepreneurial ambition from around the world. This talent inflow will later develop impactful research, attract funding, create and join companies that benefit our economy, thereby completing a positive feedback cycle.
But for this cycle to come to life, we need our spinout policy to be transparent and economically fair. Today’s spinout playbook is neither.
The universities’ deal-making problem
When a group of academics wants to commercialise an idea based on intellectual property developed at a university, their first and most important contact point is the technology transfer office (TTO). Among other things, TTOs are responsible for agreeing with would-be spinout founders on deal terms for commercialising research.
Since 1996, UC Berkeley has taken no more than 10% equity ownership in its spinouts, regardless of their involvement in the companies. In contrast, per its latest equity policy reform of 2021, Oxford takes at least a 10% equity share in its spinouts if the spinout had no support whatsoever, and 20% in almost all cases (spinout formed by salaried academics, or licensed IP, or with some financial support from the university).
This discrepancy is not an isolated one. UK spinouts are being hindered by overly bureaucratic processes and TTOs that demand onerous terms, including unfair equity shares and revenue shares (royalties) from their potential spinouts. In a report on UK university spinouts, Beauhurst tracked the equity stakes of UK universities in ~600 spinouts between 2011 and 2021. It found that universities held on average 24% of the company's equity in the year of spinning out. This was roughly consistent with our spinout.fyi data on 67 UK companies, where universities took an average 20% equity (24% excluding 0% equity deals), more than 3 times what the 16 US respondents in our data reported.
This aggregate number hides a significant disparity between UK universities. According to Beauhurst, the median university’s share of equity in spinouts is around 25% for University of Oxford (the largest spinout producer) and Imperial College London, whereas it is around 10% for University of Cambridge and University College London. Worse yet, the equity policies of some other UK universities are significantly more aggressive. For example, University of Leeds takes more than a 60% share of equity in spinouts using its IP. For the University of Nottingham, it’s 49.9%.
The reflex of some universities, particularly in the U.K., is to raise and allocate more committed capital at their spinouts… University of Leeds, Manchester and Seffield, launched a £215m fund to invest in their spinouts. Meanwhile, the average equity requirements on spinout formation from these partner universities is: 39.1% for Leeds, 32.1% for Manchester, and 21.1% for Sheffield. This approach puts the cart before the horse: venture financing follows great opportunities that are structurally sound. A founder-friendly spinout policy will unlock capital rather than the other way around.
As a result of these aggressive policies, founders are left with only 54% of their company (median) in the year of spinning out, and many find themselves minority owners. TTOs claim to put the universities’ IP to the best use, but their policies don't set the right incentives for founders and stifle academic entrepreneurship.
Low equity stakes for founders are only one side of the problem. The other is the complete lack of transparency from TTOs on spinout deal terms. University TTOs in the UK currently operate under opaque terms - we know this to be true beyond the UK too. Many don’t share why their deals are structured the way that they are. For example, Cambridge is the second largest producer of spinouts, and seems to offer reasonable equity terms to founders when compared to other UK universities. However, Cambridge founders can’t freely find what terms to expect from their TTO. We have tried to create an open index of equity policies for top UK universities using available online resources. But we quickly abandoned this initiative, as information about this was at best vague, and often unavailable. The lack of predefined terms means that negotiations over spinout deal terms can be extremely long. Spinout.fyi data shows that ⅔ of spinout deals take more than 6 months to complete, and 27% take more than a year. This is especially prohibitive for companies in domains like AI where research is very active and mostly open: founders need to be able to quickly commercialise their invention.
Nevertheless, there are some bright spots: Imperial College London and UCL both offer so-called “founder-driven tracks”: without help from the TTO to set up their company, the founders get a 95% share of equity, and 90% otherwise. These tracks have been established only recently — in 2017 for Imperial College and in 2021 for UCL. Their consequences on spinout success should be scrutinised by university leaders and government officials, as they constitute an important departure from the aggressive and opaque model which seems to be the current norm.
Greater transparency around the reality of the tech transfer process will level the playing field for founders. It will eventually encourage more entrepreneurs to set up spinouts of their own, working with universities rather than seeking alternative funding. This is why we created spinout.fyi, an open database where founders can anonymously share their spinout deal terms and give a leg up to future founders.
A policy solution?
To come up with a solution, we recently brought together a mix of representatives from different sides of the spinouts process: founders, former TTO officers, investors, and policymakers, many of whom feel similarly. We need urgent policy actions to improve spinout deal terms for founders, reduce the time to company formation, and increase founder satisfaction. The UK spinout system, and most of the European system, has an important lever to improve the spinout playbook: universities are publicly-funded. Here are the important takeaways from this session:
• Rather than optimising for total revenue from spinouts, many saw as a better north star the number of spinouts a TTO can produce. This would encourage TTOs to adopt a risk taking approach to making deals. Successful TTOs “want to make deals happen” and optimise for long term success, not short term gains. To encourage this, the government could set new KPIs for TTOs such as the number of spinouts created in a given timeframe. This would push TTOs to close more deals by offering better equity terms to founders and reducing the time from project inception to company formation.
• More public funding should be directed towards teams demonstrating both technical expertise and commercial skills and experience. This would encourage academic founders to pursue their commercial venture ideas without fear of abandoning years of academic career building.
• The government could go even further, and impose a uniform Simple Agreement to Spinout that sets a transparent standard for spinouts, which I detailed in a previous blog post.
• The UK government could also subsidise alumni ecosystems. Alumni networks are too often an afterthought in the universities’ strategy. Maintaining alumni databases, keeping alumni engaged with their schools, inviting them to events, and fostering exchange between students and alumni are all essential to maintaining the universities’ attractiveness. What’s more, an engaged alumnus will be more likely to donate back to their university. To understand why this is important, consider the fact that Oxford’s IP income in 2020 was $19M, while it received $187M in gifts and pledges in the same year. And the latter still has room to grow: Stanford received $1.1B! Keeping the university attractive and the founders happy is simply a much more important matter than maximising spinout revenue.
Freshly graduated in 2019, UC Berkeley doctors Robert Nishihara and Philipp Moritz joined forces with Professors Stoica (again!) and Jordan, and launched another spinout, Anyscale, which was valued at $1B already in 2021. This is the second billion dollar spinout coming out of RISELab. And it wouldn’t have existed if Professor Stoica was faced with an adversarial TTO the first time around, when he spun out Databricks.
To fix Europe and the UK’s spinout problem, we need our universities to be even more attractive. To put it simply: like for Databricks, we need our students and professors to talk our universities up. The best universities already work towards this outcome by providing the best coursework for students, removing the administrative burden on professors, and funding elite research labs. Why should TTOs have a different objective? TTOs need to collaborate with founders. Their goal should be to increase founder satisfaction over generations. To avoid arduous, tense negotiations, let’s set out a simple, transparent standard agreement for our spinouts that will let our academics focus on their venture and save them, and technology transfer officials, valuable time and resources.