DeepMind shakeup tests UK AI ambitions as Google shifts influence to California

The DeepMind shakeup UK AI ambitions debate has sharpened after Demis Hassabis stepped down as head of Google DeepMind, according to Bloomberg. Hassabis will become DeepMind chair and Alphabet’s chief scientist, while Google is reported to be concentrating more of its AI work in California, a shift that puts London’s influence inside one of its defining AI companies under scrutiny.

Bloomberg reported that Hassabis has relinquished oversight of Gemini, Google’s flagship model. Koray Kavukcuoglu, his successor, plans to move additional engineers west after Google transferred several senior lab leaders from London to California, Bloomberg said. Alphabet has not disclosed the scale of any further relocation in the material available, and the reporting does not establish that DeepMind is closing or reducing its London research operation.

The immediate significance is less about a street address than control. DeepMind was founded in London and acquired by Google in 2014. Google combined it with Google Brain in 2023. The lab has long served as evidence that Britain could produce frontier AI research, even as ownership and ultimate corporate decision-making sat with a US technology group.

What does the DeepMind shakeup mean for UK AI ambitions?

It raises the risk that strategic product decisions, senior technical talent and the center of gravity for Google’s AI development move further from the UK. Those shifts could affect the London startup ecosystem’s access to experienced operators and researchers, though that outcome is not yet settled. Bloomberg quoted a London-based AI consultant who described the change as potentially damaging to the UK’s reputation.

There is also a material counterweight. Bloomberg reported that Hassabis will continue to run London-based Isomorphic Labs, Alphabet’s AI drug-discovery subsidiary, and spend more time there. Isomorphic Labs, founded in 2021, has signed pharmaceutical partnerships and raised capital from outside Alphabet, but had not yet moved an AI-designed drug into clinical trials, Bloomberg reported.

The AI Now Institute has described DeepMind as both a symbol of British AI strength and an example of the country’s difficulty competing at the frontier without US patronage. That framing points to the broader test exposed by the leadership change: whether Britain can retain executive decision rights, senior researchers, model intellectual property and sufficient computing capacity, rather than relying on the continued presence of a UK-founded subsidiary.

Google DeepMind still has an operational role in Britain. In June, the company said it was working with the UK government, Google Cloud, Faculty, and planning authorities in Barnet, Dorset and Camden on a Gemini-based planning prototype. Google DeepMind said the project aimed to cut decisions on householder planning applications by 50% and could be offered nationally from 2027 after trials. Those are project targets, not results, but they show the company’s UK ties extend beyond its research lab.

For founders and policymakers, the harder question is whether those ties translate into durable domestic capability. A London office, government pilots and a high-profile chairmanship carry less weight if the people allocating compute, setting model priorities and directing commercialisation increasingly sit elsewhere.

This story draws on original reporting from Sifted.