Google lost two of its most prominent AI leaders within days of each other, both to rivals. Noam Shazeer, a vice president of engineering and co-lead of the Gemini models, said on June 18 that he is joining OpenAI.
A day later, John Jumper, the Nobel laureate who led the AlphaFold project at Google DeepMind, announced he is leaving after nearly nine years to join Anthropic. Both Google and the hiring labs confirmed the moves. The back-to-back exits are a striking signal in an intensifying contest among AI labs for a small pool of elite researchers.
Each is a marquee name. Shazeer co-authored the 2017 paper “Attention Is All You Need,” which introduced the Transformer architecture underpinning nearly every modern large language model. He had been at Google since 2000, left in 2021 to found the chatbot startup Character.AI, and returned in 2024 through a Character.AI licensing deal reported at about $2.7 billion, after which Google made him a Gemini co-lead.
Jumper shared the 2024 Nobel Prize in Chemistry with DeepMind CEO Demis Hassabis for AlphaFold, which predicted the structures of more than 200 million proteins. He said he will take time off before starting at Anthropic, where his role has not been disclosed.
The losses cut close to Google’s core. Shazeer’s architecture work sits beneath the models powering Gemini and Google’s AI search features, and Jumper’s research made DeepMind’s scientific reputation. Bloomberg reported that DeepMind staff have raised concerns the company lacks a clear enterprise product for AI coding tools, an area where Anthropic and OpenAI have momentum, and that Jumper had recently shifted toward coding work.
Hassabis publicly thanked Jumper, saying AlphaFold “lit the way for AI in science and medicine.” Alphabet shares fell about 5% to 6% on June 22, with some reports tying the drop to worries about AI spending and talent retention.
Why It Stings for Google
Google paid a high price to acquire Shazeer once and could not keep him, which sharpens the retention question. For OpenAI, his arrival adds depth in foundational architecture, consumer AI products and large-scale infrastructure as it heads toward an IPO.
For Anthropic, Jumper is a statement hire that anchors its growing push into AI for science; the company has acquired the startup Coefficient Bio, launched Claude for Life Sciences and scheduled an AI-for-science event for June 30. His move sets up a three-way scientific race with OpenAI and Google’s own Isomorphic Labs. Notably, Google retains its large equity stake in Anthropic even as it loses the researcher to it.
The Talent War
The departures fit a broader pattern of frontier researchers choosing well-funded challengers over Big Tech AI divisions. Andrej Karpathy, an OpenAI co-founder, joined Anthropic in May, and AlphaGo researcher David Silver has reportedly left DeepMind to start his own venture.
The money involved is extraordinary: OpenAI’s Sam Altman has said Meta’s signing bonuses can reach $100 million, while Anthropic’s Dario Amodei has said he will not match nine-figure offers, arguing that alignment-focused talent cannot simply be bought.
One industry report put Anthropic’s two-year retention at about 80%, the highest among frontier labs. For Google, the open question is whether it can hold onto the senior researchers who built its lead.
