Why Google DeepMind disbands its Nobel Prize-winning AlphaFold team and what it means for AI

Transfers within Alphabet

Some team members were shifted to Isomorphic Labs, Alphabet's drug-discovery company .

High-profile departures

In June 2026, John Jumper — who won the 2024 Nobel Prize in Chemistry for AlphaFold — announced he was leaving DeepMind after nearly nine years to join Anthropic . Shortly after, two other AlphaFold co-authors, Jonas Adler and Alexander Pritzel, followed him to Anthropic . In total, nearly a quarter of the original full-time DeepMind authors of the AlphaFold paper have now left the company entirely .

What the Gemini pivot means for DeepMind's scientific mission

From specialized to general

DeepMind is betting that Gemini-powered AI agents can automate and accelerate scientific discovery across biology, chemistry, physics, and materials science more effectively than one-off models like AlphaFold .

AlphaFold itself continues

The AlphaFold protein structure database (covering 200 million+ proteins) and the technology itself remain available. The move does not retire the product — it ends the dedicated team that built it .

Risk to deep-domain excellence

Critics argue that disbanding a Nobel-winning team signals a willingness to sacrifice deep scientific specialization for the race to build the most capable general-purpose AI system, potentially slowing progress in areas like drug discovery that relied on AlphaFold's focused expertise .

Broader industry trend

Talent flowing to frontier AI labs

The departure of top AlphaFold scientists to Anthropic — and the parallel loss of Gemini co-lead Noam Shazeer to OpenAI — illustrates a wider pattern. The most intense competition in AI today is around general-purpose agents and foundation models, and researchers with deep expertise in any subfield are being lured to work on those core challenges.

Resource reallocation across the sector

DeepMind's move mirrors decisions at other labs: domain-specific tools (protein folding, weather prediction, game-playing AIs) are being de-prioritized as companies concentrate compute, talent, and funding on the race to build the most capable general AI agents. A single general-purpose "scientist AI" is seen as offering a larger competitive payoff than maintaining many narrow expert systems.