Google DeepMind has officially disbanded its AlphaFold team, the group behind one of the most celebrated AI breakthroughs in scientific history. The move, which became public on July 29, marks the end of a dedicated unit that mapped protein structures for hundreds of millions of biological compounds and earned a Nobel Prize in Chemistry in 2024.

What happened and who left

Roughly 25% of the original authors on AlphaFold’s foundational research papers have departed DeepMind entirely.

The most notable exit is Nobel laureate John Jumper, whose work on AlphaFold 2 essentially solved the protein folding problem that had stumped biologists for half a century. Jumper, along with researchers Jonas Adler and Alexander Pritzel, has landed at Anthropic, the AI safety company founded by former OpenAI executives.

Researchers who stayed at DeepMind have been reassigned to other projects: enzyme design, nuclear fusion research, genomics, and Isomorphic Labs, Alphabet’s drug discovery subsidiary.

DeepMind’s new strategy centers on Gemini, its flagship large language model family, which the company wants to position as a general-purpose engine for scientific discovery. Rather than maintaining specialized teams for individual scientific problems, DeepMind is betting that a single, scalable AI system can tackle many domains simultaneously.

Why crypto markets should pay attention

First, the talent migration matters. Jumper and his colleagues moving to Anthropic strengthens a company that has been increasingly relevant to crypto infrastructure. Anthropic’s Claude models are widely used across DeFi analytics, smart contract auditing, and trading bot development.

Second, the DeSci sector, which includes tokens and protocols designed to decentralize scientific research funding and data sharing, has been building momentum precisely because centralized labs like DeepMind can make unilateral decisions like this one.

Third, there’s the biotech token angle. Several crypto projects have built tokenized ecosystems around drug discovery and protein research, leveraging AlphaFold’s publicly available database as foundational infrastructure. The disbandment of the team responsible for maintaining and advancing that database introduces uncertainty about its long-term development trajectory.

The bigger picture for AI and markets

The AlphaFold database itself remains one of the most valuable open scientific resources ever created. Its contribution to mapping protein structures has been transformative for drug development, agricultural science, and materials research.

AlphaFold earned a Nobel Prize. The team that built it is now scattered across competitors and adjacent projects. The database they created will continue to power research for years, but active development under a dedicated team has effectively ended.