John Jumper, co-creator of AlphaFold and 2024 Nobel laureate in chemistry, announced his departure from Google DeepMind on June 20 after nearly nine years. He is joining Anthropic — one of Google's primary competitors in the race for AI leadership. The move is the latest signal that competition for top researchers has become one of the defining strategic axes of the entire industry.
Key takeaways
- John Jumper spent nearly 9 years at Google DeepMind, leading work on AlphaFold among other projects
- In 2024, Jumper and DeepMind CEO Demis Hassabis received the Nobel Prize in Chemistry for AlphaFold
- In recent months Jumper worked on AI-powered developer tools — a segment where Google struggles with commercialization
- He joins Anthropic the same week Noam Shazeer left DeepMind for OpenAI
- Google DeepMind loses two researchers of strategic importance within days
AlphaFold: from academic research to Nobel
When Jumper joined Google DeepMind in 2017 — just six months after completing his PhD — AlphaFold was still in its preliminary phase. Under his leadership the system evolved from a research experiment into a tool that fundamentally changed how protein research is conducted. AlphaFold 2, presented at the 2020 CASP (Critical Assessment of Structure Prediction) competition, achieved results surpassing the previous state of the art by margins previously considered unreachable.
In 2024, the Royal Swedish Academy of Sciences awarded Jumper and Demis Hassabis the Nobel Prize in Chemistry for developing a computational method for protein structure prediction. AlphaFold 2 models the three-dimensional structure of a protein solely from its amino acid sequence — a problem biochemists had worked on for decades. The model is available as open-source software and has since accelerated hundreds of research projects worldwide, from antimalarial drug discovery to industrial enzyme engineering.
A departure in the middle of AI turbulence
The timing of Jumper's departure is not coincidental. According to Bloomberg, he had been working in recent months on AI-assisted developer tools — a segment where Google is aggressively trying to break into enterprise customers but not seeing satisfying sales results. Meanwhile Anthropic has been gaining increasing enterprise interest precisely through its developer tools, with Claude Code playing a significant role in the company's current business narrative.
Jumper himself described the departure from DeepMind as a personal rather than organizational decision. In a post on X he wrote that Hassabis "took a real chance letting him lead the AlphaFold team just six months after finishing his PhD" and that the entire DeepMind team "taught him how to do great science." He did not disclose what he will specifically work on at Anthropic.
That same week Google DeepMind lost another significant researcher: Noam Shazeer — co-author of the landmark "Attention Is All You Need" paper and co-creator of the Transformer architecture — announced he was joining OpenAI. Shazeer had returned to Google in 2024 after the company acquired Character.AI for an estimated $2.7 billion, but now leaves DeepMind again — this time for a direct competitor.
Two departures, one market signal
The parallel departures of Jumper and Shazeer are a rare phenomenon in AI, where companies work intensively to retain key researchers through long-term incentive packages. Neither disclosed specific reasons for leaving Google, but the fact of the moves — and their direction — carries significant signal value.
Anthropic is gaining Jumper at a moment when the company is dealing with a controversial export ban imposed by the Trump administration on the Claude Fable 5 model, and its market valuation and negotiating position with enterprise customers require solid support on the research side. A Nobel Prize in a new employee's CV is, in this context, more than decoration — it reinforces the company's scientific credibility at a moment when it is being challenged by external regulatory circumstances.
OpenAI, in turn, gains Shazeer — the architect of the technological foundation underlying all of generative AI. Both acquisitions demonstrate that the classic recruitment argument "work where you can do the best science" increasingly points toward startups rather than technology giants.
Why this matters
Jumper's departure is a symptom of a deeper process: the center of gravity of AI research is shifting from large corporate laboratories toward specialized firms. Google DeepMind has enormous resources and arguably the best computing infrastructure in the world — and yet it cannot retain researchers of this caliber. The reasons are several: slower decision cycles in a large organization, commercialization pressure complicating pure research, and above all the ability to move faster in a smaller, more agile structure.
For Anthropic, Jumper's value is twofold. His AlphaFold experience — working on systems operating on biological structures — may be directly useful in AI applications for science and medicine, areas the company is increasingly exploring. At the same time, the Nobel itself is a reputational signal for potential academic partners and institutional customers who care about the scientific legitimacy of their AI vendor. For the industry as a whole, this is one more piece of evidence that advantage in AI is not built solely through access to compute — it is built through people.
What's next
- Anthropic has not disclosed what Jumper will work on. The company runs active projects in biology, model interpretability, and AI safety — all three sit close to Jumper's competencies
- Google DeepMind must answer how it will retain researchers critical to its long-term strategy — particularly as competition from OpenAI and Anthropic intensifies, with both offering different working conditions and potential equity upside
- The market is watching whether coming weeks will bring further senior departures from Google — existing signals suggest this rotation at senior level is not an isolated event
