
John Jumper, the Nobel Prize-winning researcher who led the AlphaFold project at Google DeepMind, is leaving the lab to join Anthropic in San Francisco. The move is the latest high profile jump in the scramble for top AI talent and could sharpen Anthropic’s push into science and life sciences work.
As first reported by CNBC, Jumper announced the change on X and will join Anthropic’s research ranks. CNBC notes the timing lines up with other recent senior departures from Google units to rival labs, a reminder of how quickly the talent market is moving.
AlphaFold, the Nobel, and the biology play
Jumper helped build AlphaFold, the DeepMind system that changed how scientists approach protein structure. The AlphaFold Protein Structure Database now includes more than 200 million predicted structures, a resource DeepMind says has compressed research that once took years into minutes. For that work, Jumper shared the 2024 Nobel Prize in Chemistry with Demis Hassabis, according to the Nobel Prize committee.
What the companies are saying and why the timing matters
A Google DeepMind spokesperson thanked Jumper for his contributions to the lab’s scientific work, CNBC reports. Anthropic, meanwhile, is set to host a livestreamed event called “The Briefing: AI for Science” on June 30, which appears on its events page and is listed on Anthropic's site. That briefing is likely where the company’s broader science strategy comes into sharper focus.
Why San Francisco is at the center of this
Anthropic’s headcount growth and office expansion have already reshaped the Bay Area AI landscape, and local reporting has tracked the company’s recent waves of big hires and real estate moves.
The open question now is whether Jumper will steer Anthropic’s push into lab-focused products or lead an internal research group that pairs large language models with biology workflows. Either path would send a clear signal about the company’s science roadmap. Expect more clarity after Anthropic’s June 30 briefing and any public posts from Jumper himself.









