Is Hassabis Planning to Exit Google? The Possible Buffer Before His Departure

Jeff Dean announced his departure after 27 years at Google, and Demis Hassabis stepped down as DeepMind CEO to become DeepMind chairman and Alphabet chief scientist.

A report from the organization Pathfounders claims that Hassabis originally intended to leave together with Dean, but Google, fearing a stock‑price impact, persuaded him to stay and transition to a chairman role as a temporary buffer.

The article points out that Google is increasingly concentrating AI power in Mountain View: Koray Kavukcuoglu, who will succeed Hassabis in daily DeepMind operations, moved from London to California, and core Gemini model, product, and fine‑tuning work is now centered at the U.S. headquarters, signaling DeepMind’s shift from an independent research institute to a deeper integration with Google’s product ecosystem.

Another symbolic move is the reorganization of the AlphaFold team. The formerly independent AlphaFold group was dismantled; most authors were reassigned to Gemini, code‑generation, genomics, enzyme‑design, and nuclear‑fusion projects, while some joined Hassabis‑led Isomorphic Labs. Google explains this as a strategic shift from solving isolated scientific problems to leveraging a general model like Gemini to accelerate scientific discovery.

Historically, DeepMind functioned like a long‑term research lab, tackling challenges such as Go, protein folding, and other low‑revenue projects. Today, competition from OpenAI, Anthropic and others has pushed Google to focus on Gemini, agents, programming tools, and consumer‑grade products, turning AI into a fast‑paced product war.

Hassabis emphasizes that AI’s best use is to improve human health and speed up scientific breakthroughs, citing AlphaFold as the most successful example of this vision. He envisions AGI development resembling CERN’s collaborative, cautious approach, even if it delays AGI by a decade or more.

He acknowledges that the current commercial pressure has turned AI into a “intense commercial‑pressure competition,” which, while accelerating progress, is not the way he originally imagined. Nonetheless, as both a scientist and engineer, he chooses to continue responsibly advancing Gemini and AlphaFold within this competitive environment.

If Pathfounders’ information proves correct, Hassabis’s move to chairman may simply be a holding pattern before his eventual exit, marking the possible end of the “old DeepMind” era that prioritized pure scientific exploration.