DeepMind, SwiftKey, VocalIQ—Does the UK need to worry about a brain drain in artificial intelligence?
Autonomy, DeepMind, SwiftKey, VocalIQ. Each are British artificial intelligence and machine learning startups bought by US tech giants—HP, Google, Microsoft, and Apple, respectively.
And they're hardly the only ones. Alongside growing VC funding in AI, US tech firms are snapping up British-founded startups, leading to concerns that the UK is losing the best of its artificial intelligence to Silicon Valley just as it becomes a key technology.
Simon Walker, partner in corporate technology at law firm Taylor Wessing, said the sale of AI startups to US firms isn't new and doesn't look like it'll stop soon. "It is obviously disappointing that the AI cannot be retained in the UK," he said. "However, top-of-the-market AI, such as that developed by companies such as SwiftKey and DeepMind, requires huge investment and a significant platform for its use and it is only very large tech companies which have the necessary resources and platforms."
"There are just very few people in the world with the background and the experience of working on large-scale AI projects"
It's no surprise, then, that the cash is coming from Silicon Valley; the UK simply doesn't have tech giants that compare. "More than half of the UK exits [IPOs or acquisitions] in the past five years went to US investors," said Pantelis Koutroumpis, a research fellow in innovation and entrepreneurship at Imperial College London, citing stats across the tech industry. "This is expected as the driving forces in the tech scene are on the other side of the pond."
And as AI increasingly becomes core to many tech products, experts become a limited resource. "All of the big US companies are working on the subject and then looking at opportunities everywhere—not just in the UK, but everywhere in the world," said Martin Mignot, an early stage investor at Index Ventures, which invested in SwiftKey. "There are just very few people in the world with the background and the experience of working on large-scale AI projects… Large US companies need to go where those people are."
Many of the snapped-up UK firms are the fruits of research at Britain's top universities—add to the list above Evi Technologies (Amazon), Dark Blue Labs (Google), Vision Factory (also Google) that are either directly spun out of Cambridge, Oxford, or University College London, or employ staff trained there.
Universities can sometimes benefit when a technology spun out of their own labs is snapped up by keeping equity in such companies. Oxford's awkwardly-named Isis Innovations and Cambridge Enterprise are what's known as "tech transfer companies": they help commercialise innovations at universities and ensure the school is repaid via equity. They distinguish between startups, when students or staff create a company with no intellectual property belonging to the university, and spin-outs, which involve IP owned by the university.
A spokesperson for Cambridge Enterprise said the institution pulled in £15.6 million ($21.8m) from equity realisations across all subjects in 2014/15, with the takings split between relevant university departments and future investments. Oxford's Isis Innovation said it pulled in £24.58 million ($34.4m) in revenue for the school last year, with £13.6m ($19m) handed back to departments.
The UK's "brain drain" may be stemmed somewhat by the fact not every AI acquisition is catching the first flight to sunny California
While those might sound like decent money spinners, there are caveats that limit how much the universities earn from startups and spin-outs. "Most AI technology would be software in the form of copyright which is not registerable in the UK and, subject to any funders' rights, would be the property of the creator," the Cambridge Enterprise spokesperson said, adding that startups are not required to use Cambridge Enterprise to commercialise their idea but can apply for investment and support if they want it. "As a result of the University's Intellectual Property policy, the number of companies created out of the University of Cambridge is hard to track," they added. "For example, SwiftKey is an example of a startup founded by alumni where Cambridge Enterprise did not invest."
Cambridge's coffers might miss the contribution from the $250 million SwiftKey acquisition, but there's more to such successes than cash. "I think it's encouraging for undergraduates and postgraduates to see that their peers who chose to trod the unknown and risky path of technology entrepreneurship saw success at the end of the road," said Nathan Benaich, a partner at venture capital firm Playfair Capital.
And the UK's "brain drain" may be stemmed somewhat by the fact not every AI acquisition is catching the first flight to sunny California. "Interestingly, each of Dark Blue Labs, Vision Factory and VocalIQ saw teams remain in the UK: the former two joined DeepMind's London team (which insisted to remain in London as a condition to their acquisition) while the latter set up the Cambridge-based Apple Siri team," noted Benaich.
Index Ventures' Mignot suggested that could lead to investment flowing from Silicon Valley to the UK, as well as the creation of more tech and research jobs in Britain. "It's fantastic, the way that those guys are not going to move to the US," he added. "DeepMind is building an R&D team in London on AI and SwiftKey is going to be doing the same for Microsoft, so you're going to have two of the largest tech companies in the world based in London, investing massively and offering jobs in those cutting-age spaces to young talent from Cambridge, Oxford, and the other top technical universities." In other words, that keeps money and talent in the UK, regardless of where the acquiring firm's corporate headquarters are based.
"I think that's a new trend," he added. "Previously, buyers would want to bring people into their main offices and integrate acquired start-ups into the wider business, but that's been shown to not necessarily be the best way forward," he said. "It's much more useful to keep the team as it was operating before—very autonomous, but with more resources. That's what's been giving the best results." If that trend holds, the flood might be "cash in" as well as "brains out."
Canadians are also watching their AI talent get acquired by Silicon Valley
Mignot said it would be good news for the UK's own startup economy if more acquisitions such as SwiftKey and DeepMind stayed in the UK, and eventually moved on to start their own projects using what they've learned—and earned—working for the larger tech firms. "That's how ecosystems get big, that's the first layer. You need the liquidity, you need the first companies acquiring and then hiring talent, and then these guys are going to go out and start their own thing," he said. "That's how you get the ball rolling."
British startups staying in the UK after acquisition by US companies could help build that ecosystem, pulling in investment, creating good jobs, and letting local talent learn the ropes before they in turn start their own companies.
But there are those who think we need to fight the brain drain. Koutroumpis said we need to help British startups grow, giving them an alternative to acquisitions. "Taking the risk to allow these businesses to grow and turn into scale-ups will help a lot," he said, suggesting the way to battle back is more investment at earlier stages and giving them time to grow.
Of course, it's not as though American academics are ignoring AI. Benaich pulled together some stats for me, pointing out that between 2011 to 2015 there were 12 times as many AI companies in the US than the UK, which raised 60 times the capital and saw 10 times as many "exit events" such as IPOs or mergers and acquisition activity. Tech companies are investing in AI everywhere, especially at home.
And Brits aren't the only target abroad: Canadians are also watching their AI talent get acquired by Silicon Valley, with one University of Toronto professor saying the country's AI advantage is "slipping through our fingers."
In other words, we're not the only ones worried about an AI brain drain, and such concerns may simply be the natural result of having a larger, more successful rival—but if our AI experts all end up in Silicon Valley, that's only going to get worse.