Google DeepMind dismantles AlphaFold’s Nobel team

Signals Inbox·July 29, 2026·Frontier AI

Google DeepMind has broken up the dedicated team behind AlphaFold, the protein-prediction system that helped earn Demis Hassabis and John Jumper a Nobel Prize. AlphaFold is not disappearing. The bigger shift is that Google is moving from small teams solving one huge scientific problem to Gemini systems meant to work across many parts of science.

Q1What actually happened?

The Financial Times reports that Google DeepMind has disbanded the dedicated AlphaFold team. Over the past year, most of its researchers were reassigned to Gemini projects or moved to Alphabet drug-discovery company Isomorphic Labs. Roughly a quarter of the core authors reportedly left Google.

Q2Is Google shutting down AlphaFold?

No. The software, database, research, and commercial work around AlphaFold continue. What disappeared is the standalone group built around it. Google is taking the people and lessons from AlphaFold and spreading them across a wider scientific AI strategy.

Q3Why break up a Nobel-winning team?

Because AlphaFold already completed much of its original mission. It helped solve a protein-folding problem that had frustrated scientists for around 50 years, and its database now contains more than 200 million predicted protein structures. DeepMind now wants systems that can help with many scientific tasks instead of one dedicated model for one grand challenge.

Q4What is Google betting on instead?

Google is betting on Gemini as a general scientific engine. The idea is not just to predict protein shapes. It is to build AI that can read papers, write code, use research tools, form hypotheses, and help automate larger parts of an experiment. That is a much bigger target, but also a less proven one.

Q5Where did the AlphaFold researchers go?

Many stayed inside Google and moved to Gemini or Isomorphic Labs. Others left. AlphaFold co-creator and Nobel laureate John Jumper is joining Anthropic, alongside former colleagues including Jonas Adler and Alexander Pritzel. So this is also a talent-war story: Anthropic is hiring people behind Google’s clearest scientific AI win.

Q6Does this make strategic sense?

Potentially. A dedicated team can move faster on one hard problem, while a shared Gemini platform can spread new capabilities across biology, chemistry, mathematics, and coding. The risk is that Google weakens the specialist culture that made AlphaFold work before its broader AI scientist is ready to replace it.

Q7So what should we watch now?

Watch whether Gemini produces another measurable scientific breakthrough, not just better demos. Also watch where the former AlphaFold researchers land and what they build. If Anthropic turns that talent into a strong science product before Google does, breaking up the team may look less like smart consolidation and more like an expensive talent leak.