DeepMind dissolves its AlphaFold team

Google DeepMind has broken up the dedicated team behind AlphaFold, its Nobel Prize-winning protein-structure model, and confirmed the move after the Financial Times first reported it. Most of the original AlphaFold paper’s authors have spent the past year moving to other projects, chiefly ones built around Gemini, along with enzyme design, nuclear fusion, and genomics; nearly a quarter of the paper’s full-time DeepMind authors have left the company outright. Some moved to Isomorphic Labs, the Alphabet drug-discovery subsidiary DeepMind spun out in 2021.

The departures reach the project’s inner circle. John Jumper, who shared the 2024 Nobel Prize in Chemistry with DeepMind CEO Demis Hassabis for AlphaFold, left for Anthropic earlier this summer; fellow AlphaFold researchers Jonas Adler and Alexander Pritzel are following him there. Pushmeet Kohli, DeepMind’s VP of research, told the FT that the lab’s nine-year strategy of “a concrete goal every project is focused on” has “evolved” toward Gemini-powered systems meant to assist and eventually automate parts of scientific research, rather than dedicated teams per grand challenge. DeepMind said it remains “incredibly proud” of AlphaFold’s scientific impact; the model and its public database are not being discontinued.

What it means for operators

This is a talent and strategy story as much as a science one: the lab that produced the most-cited scientific AI result of the decade is folding its authors into general-purpose Gemini work, and losing some of them to Anthropic in the process. If your roadmap depends on a specific lab’s specialized research effort continuing to exist in its current form, whether that’s AlphaFold-style science models or anything similarly narrow, this is a reminder that internal reorganizations and competitor poaching can move faster than the product surface tells you.