Google DeepMind has disbanded the dedicated AlphaFold team that won the Nobel Prize in Chemistry, with nearly a quarter of original authors leaving. Core members John Jumper, Jonas Adler, and Alexander Pritzel have joined Anthropic. Google says it will shift toward a Gemini-powered general science system rather than organizing teams around single problems.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
Google DeepMind has disbanded the dedicated AlphaFold team that won the Nobel Prize in Chemistry, with nearly a quarter of the original authors leaving. Lead John Jumper and other core members have joined Anthropic. Google says it will pivot to a Gemini-centric general science system, but the move raises concerns about long-term research investment.
On July 29, 2026, the Financial Times reported that Google DeepMind had dismantled the dedicated team responsible for developing AlphaFold. Over the past year, most of the original authors of the AlphaFold paper have been reassigned to other projects, some moving to Gemini, AI coding, genomics, enzyme design, and nuclear fusion research, while others entered Isomorphic Labs, Alphabet's drug discovery company. Google DeepMind confirmed the personnel changes to the Financial Times.
Nearly a quarter of the original AlphaFold paper authors who were full-time Google DeepMind researchers have now left the company entirely. AlphaFold lead John Jumper, researchers Jonas Adler, and Alexander Pritzel have all decided to join Anthropic. Before leaving, Jumper and Adler were reassigned to Google's internal 'Code Strike' team, tasked with improving Gemini's coding capabilities. Pritzel, an early core developer of AlphaFold, has been involved in Gemini pre-training in recent years.
A DeepMind insider described Jumper, Adler, and Pritzel as 'key core members' of the company, saying their departures caused surprise and shock internally. Jumper is the most symbolic figure—in 2024, he shared the Nobel Prize in Chemistry with Google DeepMind CEO Demis Hassabis. In June, Jumper announced the end of his nearly nine-year tenure at DeepMind to join Anthropic. In his farewell statement, he thanked Hassabis for letting him lead the AlphaFold team but did not publicly explain why he left or what project he would work on at Anthropic.
Regarding this restructuring, Google DeepMind's explanation is not about cutting scientific research but changing how research is organized. Research Vice President Pushmeet Kohli told the Financial Times that the company's strategy over the past nine years was to select major scientific challenges with clear goals and form dedicated teams around each problem; now, that strategy has evolved. The new direction is no longer about developing separate models for protein folding, weather prediction, or nuclear fusion, but rather building a general science system powered by Gemini that helps scientists retrieve information, generate hypotheses, write and run code, and gradually automate parts of the research process.
This change is already reflected in Google's recent products. In May 2026, Google DeepMind launched Co-Scientist, a Gemini-based multi-agent system that works by having multiple agents repeatedly propose, critique, and refine scientific hypotheses, applied to research on antimicrobial drugs, plant immunity, and liver fibrosis. Google is also extending large model capabilities to mathematics, physics, computer science, and algorithm discovery through systems like Gemini Deep Think and AlphaEvolve.
In other words, Google is not preparing to exit AI for Science but aims to integrate it into the Gemini foundational model and agent platform. AlphaFold represented 'one team solving one big problem,' while the new route is 'one general model serving multiple scientific domains.'
After the news broke, the tech community first questioned why Google reassigned a scientist like Jumper to an AI coding team. Reddit user Recent_Fox4339 sarcastically commented: 'They actually moved that John Jumper to improve Gemini's coding abilities? Then it's no surprise he left.' The user noted that DeepMind's direction has clearly changed in recent months and began questioning whether the company would return to the frontier scientific research that once made it famous.
Another user, TorturedPoet30, responded that Jumper's departure initially seemed sudden but now makes sense: 'Having someone with Jumper's expertise work on coding tools seems like a waste.' They also questioned why Google didn't move Jumper to Isomorphic Labs, which continues biological and pharmaceutical research. User Fit-Dentist6093 was more pointed: 'Imagine you win a Nobel Prize and are a chief research scientist, then the company tells you to do data cleaning and sorting, teaching models to become software engineering tools.'
Some users see this restructuring as a sign of DeepMind's increasing corporatization. Recent_Fox4339 said Hassabis now appears more like a Google executive than the research lead they remember. Other comments pointed to commercial competition. User mintaka wrote: 'Seems like this whole AGI thing is ultimately about money.' User IntelligentMedium698 argued that current AI companies prioritize growth and competition over user needs, and Google's intense focus on coding agents is part of the industry-wide race for frontier models.
On Reddit, another perspective holds that AlphaFold has not been abandoned but moved to Isomorphic Labs. User Aaco0638 pointed out that biomedical research around AlphaFold continues at Isomorphic Labs, so describing the restructuring as 'Google abandoning AlphaFold' is inaccurate. Another user, swimmingupclose, also believes that Isomorphic Labs has in many ways taken over the original AlphaFold team's work, and the team split itself may not be a wrong decision.
Isomorphic Labs was founded by Alphabet in 2021 with the goal of applying AlphaFold's breakthroughs in structural biology to drug discovery. Currently, the company is developing drug design systems that go beyond mere structure prediction, described as a platform extending from AlphaFold 3 to real drug development pipelines.
Thus, some argue that this restructuring cannot be simply summarized as 'Google abandoning AI for Science.' The data, models, and research methods left by AlphaFold remain operational, and biomedical commercialization is handled by Isomorphic Labs. What has truly changed is the allocation of research resources within DeepMind: fixed teams are shrinking, and scientists are being dispersed into Gemini, scientific agents, and other interdisciplinary projects.
However, the community's concerns are not unfounded. AlphaFold's success came precisely from long-term, focused, dedicated investment: researchers tackled a clear but high-risk problem persistently, without first considering whether the model could become a platform, gateway, or commercial product. Now that these researchers are being pulled into Gemini and AI coding competition, the question arises whether DeepMind can still reserve space for the next 'AlphaFold-level project' that requires years of investment with no short-term monetization. In this sense, this is not the end of AlphaFold's technical life, but it may be the end of an era. AlphaFold has not disappeared, but the organizational form that nurtured it is fading away.
This article is primarily based on a Financial Times report and Reddit community discussions. Core facts (team disbandment, personnel departures, new direction) were confirmed by Google DeepMind to the Financial Times and are considered confirmed. Community views and comments are source claims and do not represent facts.
The dedicated AlphaFold team has been disbanded, core members have moved to Anthropic, and Google is pivoting to a Gemini-centric general science system. AlphaFold technology remains operational, but the organizational form that created it has changed, raising concerns about long-term research investment.
Primary report
Primary source
Also covered by 4 sources