Britain’s got AI talent but that’s not enough

It’s the kind of crazily ambitious investment that seemingly only happens in Silicon Valley near the top of a hype cycle: a pre-product, lossmaking artificial intelligence start-up raises $1.05bn from SoftBank, Microsoft and Nvidia to fund a dizzying ambition for global domination. In fact, this deal was sealed in London last week when Wayve, which develops software for autonomous vehicles, announced Europe’s largest AI start-up fundraising.

If any British technology company is to match chancellor Jeremy Hunt’s desire to create a homegrown Microsoft over the next decade worth more than $1tn, then Wayve is probably as good a bet as any. As a global pioneer in the white-hot field of embodied AI — interacting and learning from the environment — Wayve could sell the software for millions of self-driving cars if it spends its money wisely and executes its plans impeccably. “Embodied AI is going to be a sector that produces trillion-dollar companies,” says Alex Kendall, Wayve’s co-founder and chief executive. “And certainly our ambition is to be one of those.”

The seven-year history of Wayve highlights the strengths of the UK tech industry: foundational research from a world-leading university; early financing from a vibrant start-up funding community; and a flexible post-Brexit regulatory regime that accommodates new technology. But Wayve’s latest funding round also exposes one of London’s lingering flaws: the growth capital needed to accelerate the company’s expansion has come from abroad. If Wayve delivers on its promise, it will now be foreign investors who benefit most.

That may not matter much to Wayve, which has already opened an office in Silicon Valley and outgrown the UK, as indeed it must if it wants to be a global industry leader. “Our roots in Cambridge and London are very important to me personally. But we are a global company building a global product,” the New Zealand-born Kendall, who was a star academic at Cambridge university before founding Wayve, tells me in a video interview from Seattle.

It has been a good month for London’s reputation as a world AI hotspot. Last week, two fast-growing US companies — Scale AI, a training data company, and CoreWeave, an AI-focused cloud computing provider — announced they were opening their European headquarters in London. In addition, CoreWeave is investing $1bn in data centres in the UK. Hunt has also been busy plugging Britain’s merits to tech leaders at the chancellor’s Buckinghamshire estate at Dorneywood this week.

Britain boasts several top-ranked universities churning out AI researchers as well as the global talent magnet that is Google DeepMind, which ranks among the strongest industrial AI labs in the world with more than 2,000 employees. “At the end of the day, talent still matters most in this space. London is incredibly important,” says Jordan Jacobs, managing partner of Radical Ventures, a specialist Toronto-based AI investment fund. 

But before British politicians indulge in too much congratulatory backslapping, it is worth noting how fast the AI industry is developing in the rest of the world, most notably in China, France and the United Arab Emirates, and questioning how far the benefits of the AI revolution are being diffused through the rest of the economy. 

At the top of the AI food chain, the giant US tech companies still dominate. According to the McKinsey Global Institute, the Magnificent Seven AI-obsessed US tech companies — Microsoft, Alphabet, Apple, Amazon, Meta, Nvidia and Tesla — spent $200bn on research and development last year — about half the total equivalent public and private sector spending in Europe. US start-ups also attracted 73 per cent of the $42.5bn invested globally in the AI sector last year, according to data firm CB Insights. In such a US-dominated world, Britain is at risk of becoming little more than a research offshoot.

As a recent paper published by the AI Now Institute noted, Britain’s much-trumpeted success in AI across a narrow range of metrics has obscured a spasmodic policy framework, poor computing infrastructure and a squandering of public sector data, invaluable for training AI models. Since 2010, Britain has announced 11 different growth plans or industrial strategies, none of which has spurred significantly higher investment. Britain accounts for just 1.4 per cent of total global computing capacity, behind Italy, Russia and Finland.

Britain certainly has good reason to applaud its ambitious AI researchers and entrepreneurs. But the country desperately needs to deepen and broaden adoption of the technology if it is to reap its full benefits.

john.thornhill@ft.com