Google Shifts AI Power to California in Race Against Anthropic, OpenAI

• Google is centralising its AI leadership in Mountain View, California, to better compete with Anthropic and OpenAI and accelerate development of its Gemini AI model.
• Key AI leaders like Koray Kavukcuoglu have moved to California, while DeepMind chief executive officer Demis Hassabis has shifted to a strategic chairman role, with some London staff leaving.
• Google faces pressure after delays and talent departures, prompting a $205 billion investment in 2026 and a focus on delivering results quickly under Kavukcuoglu's leadership.

SAN FRANCISCO - Alphabet’s Google is concentrating its artificial intelligence leadership at its Mountain View, California, headquarters to gain momentum in the accelerating race against Anthropic and OpenAI to build the world’s dominant AI models.

The tech giant appointed Koray Kavukcuoglu, who moved to Mountain View in 2026, to run its sprawling AI research and operations, while Demis Hassabis, DeepMind’s co-founder and long-time chief executive officer, has stepped back from day-to-day operations to become chairman of Google DeepMind and Alphabet’s Chief Scientist.

Sebastian Borgeaud, the head of a key AI coding effort, also moved from the UK to California. Others have departed from the company’s London headquarters altogether.

Efforts to centralise Google’s AI leadership are aimed at reversing a trend that has beset Google since at least 2023, when it merged two prized labs that were operating independently: Google Brain, which was housed in the company’s Mountain View headquarters, and DeepMind, which is rooted in London.

While the labs joined under the banner of Google DeepMind, researchers continued to operate on separate continents.

This complicated decision-making and frustrated talent on both continents, according to people familiar with the matter. 

Kavukcuoglu is formally assuming responsibility for Google’s AI ambitions as it loses ground to rivals. The company is months behind schedule in launching the most powerful version of its flagship AI model, and many see its products as an afterthought in the market for coding automation, which has emerged as one of the most lucrative early applications of AI.

Researchers’ frustration with Google’s competitive position has mounted, prompting a wave departures to launch startups – including legendary Google engineer Jeff Dean – and defections to rivals Anthropic and OpenAI, which offer a place at the cutting edge of the market and the riches of an upcoming initial public offering.

And Google has pledged to spend as much as $205 billion (S$262.5 billion) in 2026 alone on capital expenditures to support its AI ambitions, pressuring Kavukcuoglu to show he can deliver results quickly. 

Some of the company’s departures include veterans from DeepMind’s London office.

David Silver, a distinguished researcher there, left earlier in 2026 to form a new startup, Ineffable Intelligence, that raised $1.1 billion in seed capital.

Simon Bouton, DeepMind’s Chief Experience Officer, a 13-year veteran who oversees operational staff, is also planning to leave for CuspAI, a London-based startup, according to a person familiar with the matter. Bouton did not reply to a message for comment.

Google said London remains an important hub for Google’s AI talent and ambitions. The company recently unveiled a new office in London, where Hassabis will remain and focus on AI vision and strategy.

The moves over 2026 have meant that many of Google’s key AI leaders who are focused on driving forward the Gemini model – which is the linchpin of its AI strategy – are united in one place. 

Since the AI boom kicked off, Google

Over time, oversight of Gemini shifted to Kavukcuoglu. 

The new arrangement has also been reflected in how Brin describes his own role inside Google’s AI efforts.

In June, at an event for AI researchers and industry leaders, Brin was asked how he split responsibilities with Kavukcuoglu and Hassabis. Brin described himself as deeply involved with Gemini and working closely with Kavukcuoglu.

“I would say, I poke and prod him and the team about, ‘Hey, are you really doing that?’ Sometimes a little bit disruptive, not going to lie,” Brin said.

“But Koray actually organizes the groups and has them deliver stuff.”

A longtime DeepMind leader who founded the deep learning team and pioneered several major research breakthroughs at the company, Kavukcuoglu has kept a relatively low profile compared to Hassabis, but has been quietly rising through the ranks in DeepMind’s leadership.

Last summer, Kavukcuoglu was promoted to be Google’s Chief AI Architect.

Yann LeCun, a pioneer of deep learning who mentored Kavukcuoglu while he was a PhD student at New York University, described him as “a great engineer” who “developed excellent research management skills” during his rise at DeepMind.

LeCun added that Kavukcuoglu’s early research helped convince people that AI could fundamentally change the way machines process language.

Now, he will have his work cut out for him as Google deals with the fallout from the departures of many of its most prominent AI leaders.

Kavukcuoglu is the only Gemini co-lead remaining at the company. BLOOMBERG