Nobel laureate Jumper joins Anthropic

John Jumper announced on June 19 that he is leaving Google DeepMind for Anthropic. He had been at DeepMind since 2017, most recently as a vice president and engineering fellow, and in 2024 he shared the Nobel Prize in chemistry with DeepMind chief Demis Hassabis for AlphaFold, the system that predicts a protein’s three-dimensional shape from its amino-acid sequence and has now mapped more than 200 million of them.

He plans to take time off before starting, and has not said what he will work on at Anthropic. The move is being read across the industry as one of the year’s most significant hires, a marker in the talent fight between frontier labs.

Update, June 26: the exits are not slowing. TechCrunch reported on June 24 that Jonas Adler and Alexander Pritzel, two researchers who played key roles in building Google’s Gemini model, are also leaving for Anthropic. That puts several senior departures from Google in the space of a week, against a backdrop of OpenAI and Anthropic preparing to go public and dangling richer equity.

What it actually tells an operator

Not much you can act on this week, and that is the honest read. One researcher changing employers does not change which model writes your emails or your code tomorrow.

What it does signal is where deep scientific-AI talent is heading. AlphaFold is the clearest case of AI producing a durable result outside chatbots and image generators, and the person most associated with it is now at Anthropic. For operators that is a data point on a lab’s direction, the kind worth noting and then watching, not the kind worth re-platforming over. Treat a marquee hire the same way you should treat a marquee benchmark: interesting, not yet proof of anything you can use. The thing that should move your stack is a tool that does your real work better, not a headline about who joined whom.