Hassabis steps back as Google DeepMind CEO in AI shake-up

Alphabet has confirmed that its co-founder and figurehead, Demis Hassabis is moving into the chairman’s seat at the AI lab, adding the title of chief scientist at parent company Alphabet, while keeping his post leading Isomorphic Labs, the drug-discovery venture he also founded.

Koray Kavukcuoglu, previously DeepMind’s chief technology officer, now runs daily operations as senior vice president, reporting directly to Alphabet chief executive Sundar Pichai rather than holding a standalone CEO title.

The reshuffle did not happen in isolation. Alongside it, Alphabet confirmed that Jeff Dean, a 27-year Google veteran and DeepMind’s chief scientist, is leaving to launch an independent venture called Discovery Loop, alongside senior fellow Sanjay Ghemawat, DeepMind vice president Oriol Vinyals and Google Brain co-founder Quoc Le. Google will invest in the new outfit and supply its cloud infrastructure, keeping the departing talent inside its commercial orbit even as they walk out the door.

Why the market flinched

Investors did not take the news calmly. Alphabet’s shares dropped more than 4% in the hours after the announcement, wiping out a chunk of a rally that had seen the stock climb 13% since its last earnings report. That reaction speaks to genuine unease rather than routine reshuffle jitters. Axios reported that Google’s next flagship model, Gemini 3.5 Pro, is running months behind schedule, with sources pointing partly to sagging morale inside DeepMind following a string of senior departures, including one of Gemini’s co-leads to a rival lab.

Gemini’s release cadence has become a proxy for how aggressively Google Cloud pushes new AI-optimised infrastructure into its regions, and any wobble in the model roadmap tends to ripple into capacity planning conversations with hyperscale customers.

Hassabis, who shared the 2024 Nobel Prize in Chemistry for his work on AlphaFold, framed the move in distinctly grand terms in a note to staff. “I’ve been working towards AGI my whole life, and now, like many of you, I feel it is close at hand,” he wrote.

“It’s critical that we collectively get the next steps right to ensure this all goes well for humanity and we usher in an incredible new age of discovery and wonder.” He added that he wanted “the time and space to focus on the big picture and help influence what is to come” rather than continue managing DeepMind’s day-to-day operations.

Pichai’s own message leaned harder on competitive urgency. “We’ve got amazing talent, world-class compute, and products that bring AI to more people than any other company,” he said, adding that Google needs to “accelerate all this work and stay focused on the AI frontier.” Kavukcuoglu, for his part, struck a tone aimed at reassuring anxious staff and partners that the lab’s course would hold steady. “Our mission to build AI responsibly to benefit humanity does not change,” he said. “My primary focus is ensuring we have the clearest possible path to delivering on our ambitious Gemini roadmap as we progress toward AGI. In today’s rapidly evolving landscape, we must operate with absolute intention and velocity.”

Dean, meanwhile, gave the New York Times a rather more candid explanation for why he was leaving a public company to strike out on his own. “We might make decisions that are not necessarily in the company’s purist financial interests,” he told the paper, a comment that reads as a pointed nod to the tension between scientific ambition and shareholder pressure that has been building inside Google’s AI division for some time.

What it means for infrastructure buyers

None of this changes Google’s underlying commitment to AI capacity, at least not on the evidence so far. The company has spent the past year on a global building spree: a $15 billion data centre commitment in Andhra Pradesh, India; a new Google Cloud facility in Winschoten, in the Netherlands, explicitly framed by local Google executive Marco Ynema as support for the country’s “AI ambitions”; a long-delayed but now confirmed Austrian site; and a £5 billion UK investment programme that included the Waltham Cross data centre opened earlier this year. DeepMind itself is also midway through building an automated research lab in the UK, a project the government has held up as proof of transatlantic tech cooperation.

Google is not alone in facing this scrutiny. OpenAI continues to burn cash while making multi-billion-dollar infrastructure commitments, and Microsoft has fielded its own questions about overreliance on a single AI partner. What sets Google apart is that its AI research arm and its cloud infrastructure business are now more tightly bound together under one reporting line than at any point in DeepMind’s history.

For an industry still trying to work out how much of the current data centre build-out is durable demand and how much is speculative positioning, a shake-up at the top of one of the field’s most closely watched labs is worth more than a passing glance.

Hassabis stepping back from operational control does not signal retreat from AI investment, quite the opposite, if his own language about AGI being “close at hand” is any guide. But it does suggest that the path to get there inside Google is being redrawn in real time, and that the executives betting billions on the infrastructure to support it will want to watch the next few quarters closely.

Google threatens to overtake OpenAI in AI race with investment push

Google powers up Netherlands data centre to boost AI economy

Google finally commits to much-anticipated Austria data centre