Skip to content
News Leadership

Nobel Laureate John Jumper Leaves Google DeepMind for Anthropic After Nine Years

· by Pondero Newsdesk

The short version

John Jumper, who co-led AlphaFold 2 and shared the 2024 Nobel Prize in Chemistry, announced on June 20 that he is joining Anthropic. The move follows Noam Shazeer's departure to OpenAI by one day, and strips Google of two of its most decorated AI scientists in under 48 hours.

Nobel Laureate John Jumper Leaves Google DeepMind for Anthropic After Nine Years

John Jumper, the computational chemist who led the AlphaFold 2 project and shared the 2024 Nobel Prize in Chemistry with Google DeepMind CEO Demis Hassabis, announced on June 20 that he is leaving DeepMind after nearly nine years to join Anthropic. His role at Anthropic has not been disclosed.

What happened

Jumper posted the announcement on X, writing that Hassabis "took a real chance letting me lead the AlphaFold team just six months after finishing my PhD," and that "GDM is a special place," per TechCrunch's reporting on his post. He did not specify his new responsibilities at Anthropic. Anthropic has not commented publicly, and no role title has been announced.

Bloomberg separately reported that in his final period at DeepMind, Jumper had shifted from biology to Google's internal coding tools program, a product the company has had difficulty selling to enterprise customers.

The announcement arrived one day after Noam Shazeer, a co-author of the 2017 "Attention Is All You Need" paper and a co-lead of Google's Gemini model program, announced his own departure to OpenAI. Both departures occurred in the same 48-hour window, taking Google's two most publicly recognized AI scientists to its two principal competitors.

Hassabis responded on X by writing that AlphaFold "changed the world" and demonstrated what AI could achieve for science and medicine.

Why AlphaFold matters

To understand the weight of Jumper's move, it helps to look at what AlphaFold 2 actually did. Predicting the three-dimensional structure of a protein from its amino acid sequence had been an open problem in biology since 1961. At the CASP14 competition in 2020, AlphaFold 2 achieved accuracy comparable to experimental laboratory methods, a result the judges described as a solution to the decades-old challenge. Jumper and his team published the full architecture in Nature in July 2021.

DeepMind subsequently released a database of predictions covering more than 200 million proteins. Per TechTimes, that database is now used by more than 2 million researchers across 190 countries and has accelerated drug discovery and vaccine development at a scale no prior computational tool in life sciences had reached. The 2024 Nobel committee awarded Jumper and Hassabis one half of the chemistry prize specifically for this work. Jumper is the youngest chemistry laureate in more than 70 years, per TechTimes.

The technical core of AlphaFold 2 is an architecture called the Evoformer: 48 stacked attention blocks that process evolutionary sequence data and pairwise residue interactions simultaneously. The attention mechanism captures long-range dependencies across the full protein chain, which matters because residues distant in a sequence can sit adjacent in three-dimensional space. That insight is why the system works where earlier convolutional approaches could not.

What Anthropic has been building toward

Anthropic did not recruit Jumper into a blank-slate organization. Throughout 2026, the company opened wet labs, published research on AI agents for biological workflows, and announced partnerships with the Allen Institute and the Howard Hughes Medical Institute's Janelia Research Campus. Those partnerships deploy Claude-based AI agents directly into single-cell genomics, connectomics, and imaging pipelines, per TechTimes's reporting on the company's science strategy. Anthropic CEO Dario Amodei has written publicly that AI-enabled biology could compress 50 to 100 years of scientific progress into five to ten years.

Jumper is not the first prominent research hire Anthropic has made in 2026. Andrej Karpathy, an OpenAI founding member and influential AI researcher, joined Anthropic's pre-training team in May 2026, per TechTimes. SignalFire's 2025 State of Talent Report found that DeepMind engineers were nearly 11 times more likely to leave for Anthropic than the reverse.

Anthropic is hosting an AI-for-science event on June 30. That event is the earliest plausible signal of what Jumper's formal role will look like.

Context: Google's week in talent

The back-to-back departures have a structural dimension that goes beyond personnel losses. Shazeer co-invented the Transformer architecture that underlies every major large language model in use today. Jumper built the most consequential application of AI to science this decade. Google had spent approximately $2.7 billion in August 2024 to bring Shazeer back from Character.AI, yet he departed for OpenAI in under two years.

Google enforces noncompete clauses of six to twelve months for UK-based researchers, in some cases placing staff on full-pay garden leave to delay transitions to competitors, per TechTimes. DeepMind retains substantial research capacity and continues publishing at the frontier of AI. The departures do not diminish AlphaFold's existing database or Google's operational footprint. The signal they send is narrower: whether the most exceptional researchers believe the most important work happens inside Google's walls.

For Anthropic, Jumper's credentials resolve a specific gap. The company has a strong commercial trajectory and a safety-focused brand, but its scientific identity has rested primarily on language model work. A Nobel laureate who has already demonstrated that AI can solve a 60-year-old biology problem carries a credential no benchmark score can replicate.

What to watch next

Anthropic's June 30 science event is the next concrete milestone. If Jumper appears or is named in that event, it will confirm his position within the company's science or biology division rather than in general model research. The question of whether other DeepMind researchers follow, particularly those who worked closely with Jumper on the AlphaFold team, is the second thing to track. Google naming a successor to Hassabis's science research leadership structure, or any internal reorganization at DeepMind following both departures, would be a third signal worth watching in the weeks ahead.

Sources