Here are two questions for venture capital: how do you spot potential in a sea of AI start-ups? And how do you build a university and enterprise system that really can create the next Google?
Patrick Chung, co-founder and managing general partner of Xfund, is well qualified to answer both. Patrick wrote Open AI founder Sam Altman’s first cheque and established Xfund – a VC firm that backs founders from US universities. His previous investments include personal genomics company 23andMe and AI pioneer Kensho.
On a flying visit to the UK – where Xfund is a partner to Future Planet Capital – he has straightforward answers when asked how he identifies value. “We chase the talent,” he says. “We let the talent decide where to go.”
He has helpful insight for investors and, perhaps, for a new UK government that arrives in office having pledged to work with universities to support spinouts.
Because for years, talent-spotting has been Xfund’s approach, finding fresh ideas among the students and alumni of institutions like Harvard and Stanford.
There are plenty of successful new AI-driven companies to come, Patrick says, even if another Open AI won’t burst onto the market any time soon. No one is going to compete with the established large language models (LLMs). They’re simply too expensive to build and operate. But start-ups can licence proprietary data to develop better trained LLMs.
Take Open Evidence, an Xfund portfolio company, which has secured data from the Mayo Clinic. Its AI – trained on resources from this world-famous medical research centre – will be better able to understand medical conditions than any LLM relying on information it can find freely-available online.
Or consider Delphi, another Xfund investment, which is creating digital clones. If – with your permission – an AI engine could read all your messages and observe all your work decisions it could eventually make the same decisions that you would. This digital clone could not only record a meeting you missed, but draw the same conclusions that you would have done.
There’s a reason, Patrick explains, that some US universities do so well at incubating businesses like these. It stems from a mix of culture, understanding and knowhow and it’s an approach the UK – and some institutions across the US - could yet copy.
He identifies three core qualities that drive entrepreneurial success on campus.
The first is cultural. Turning academic discovery into commercial success is celebrated in West Coast learning. That’s still not the case in every university where – for all the excitement about spin-outs – making money can still be seen as unworthy of an academic.
“We need a culture change so this commercialisation and mass distribution is an acceptable manifestation of your discovery if you’re a lead investigator,” says Patrick.
Secondly, universities need informed, sophisticated departments to licence their intellectual property for academics who want to build start-ups. Some academic administrators can still seem baffled by business.
In one case, Patrick says, a founder wanted to gift equity from his company to his university. It was an act of gratitude towards the institution that had inspired him and trained his first staff. That founder wasn’t thanked; instead, the university demanded a payment for accepting the gift. Its administrators simply confused equity with liability.
An understanding gap remains.
Finally, universities need to be places where you can get involved in a new business, but also get back out of it again.
Tenured professors would be much more likely to make the leap from academia to business if they knew they could leap back the other way afterwards. Investors should be able to invest in businesses that have strong leadership already in place and not have to piece executive teams together for themselves, then hang around forever to run their companies.
It all means that we can still do better at commercialising academic work. For now great ideas still languish in labs for want of the right structures and leaders to set them free, Patrick says. For the VCs though, that means there are still new ideas, and more Sam Altmans, to fund.