Nobel laureate Jumper joins Anthropic as DeepMind loses Transformer co-author Shazeer to OpenAI
John Jumper, the 2024 Nobel Prize winner who led AlphaFold development, is joining Anthropic from Google DeepMind. His departure follows other high-profile exits, including Transformer co-author Noam Shazeer heading to OpenAI.
John Jumper, the 2024 Nobel Prize winner in Chemistry who led the development of AlphaFold at Google DeepMind, is joining rival Anthropic. The move marks one of the most significant talent shifts in AI research this year, signaling intensifying competition for top-tier researchers among frontier labs.
Jumper is not departing alone. Noam Shazeer, a Google veteran since 2000 and co-author of the landmark 2017 "Attention Is All You Need" paper that introduced the Transformer architecture, is reportedly heading to OpenAI. Shazeer's career path is notable: he left Google in 2021 to co-found Character.AI, then returned to Google last year when the company acquired Character.AI's technology and rehired much of its team. His move to OpenAI represents another defection from DeepMind's ranks.
The timing underscores structural shifts in AI talent allocation. DeepMind, acquired by Google in 2014, has long served as the company's flagship research arm. Last year's merger with Google Brain was intended to consolidate Google's AI efforts, yet the lab is now losing researchers to direct competitors. Jumper's Nobel recognition—shared with DeepMind co-founder Demis Hassabis—makes his exit particularly visible. Anthropic, by contrast, has been aggressively recruiting as it scales Claude and competes directly with OpenAI and Google on frontier models.
Whether these exits reflect internal friction at DeepMind or simply the market reality of AI talent scarcity remains unclear. Anthropic's ability to recruit a Nobel laureate from a competitor suggests its research culture and equity packages are now competitive with legacy tech labs. Watch for whether additional senior researchers follow in the coming months—a pattern that could reshape the competitive landscape among frontier AI labs.
