Google DeepMind has reportedly disbanded the team behind its Nobel Prize-winning AlphaFold system, signaling a sweeping overhaul of the scientific research strategy that established the London-based lab as one of the world's premier artificial intelligence laboratories. The move reframes DeepMind's priorities around its Gemini large language models and the fiercely competitive race to build frontier AI agents. For broader context on the corporate realignments reshaping the field, follow our AI industry coverage.
The reorganization was detailed in a report published Wednesday, July 29, 2026 by the Financial Times, which framed it as Google overhauling a research approach that made DeepMind one of the top AI labs in the world. According to the FT's analysis of recent job moves and sources familiar with the matter, most of the original authors of the AlphaFold papers have been reassigned over the last year.
From Grand Scientific Challenges to Gemini
The company confirmed that the affected researchers have moved to projects tied to Google's Gemini large language model, along with adjacent scientific areas including enzyme design, nuclear fusion, and genomics. Others have moved to Isomorphic Labs, the DeepMind spinout focused on AI-driven drug discovery. The FT found that almost a quarter of the full-time Google DeepMind authors of the original AlphaFold papers have left the company altogether.
Pushmeet Kohli, vice-president of research at Google DeepMind and the founder and head of its AI for Science team, told the FT that the lab's focus is shifting. "Our strategy over the last nine years has been to focus on grand challenges, a concrete goal every project is focused on," Kohli said. "The strategy has evolved."
Rather than concentrating on a single high-profile scientific problem the way AlphaFold targeted protein structure prediction, DeepMind is now also building Gemini-powered systems designed to help scientists and, eventually, automate parts of the scientific process. The pivot places the lab in direct competition with OpenAI and Anthropic to develop frontier AI agents capable of autonomous reasoning and task execution.
A Nobel-Winning Exodus
The timing underscores the magnitude of the strategic shift. AlphaFold, which predicts protein structures with accuracy that transformed biology and drug discovery, was central to the 2024 Nobel Prize in Chemistry awarded to DeepMind's Demis Hassabis and John Jumper, alongside the University of Washington's David Baker. Weeks before the FT report, Jumper, one of the Nobel-winning scientists behind AlphaFold, announced he was leaving DeepMind to join rival Anthropic.
The departures reflect a broader talent realignment across the AI industry, where researchers who once pursued long-horizon scientific breakthroughs are increasingly pulled toward the consumer and enterprise agent race. DeepMind's decision to fold its prize scientific team into the Gemini effort suggests Google sees the largest near-term value, and competitive risk, in general-purpose language models rather than in dedicated scientific moonshots.
The reshuffle comes as workplace adoption of AI accelerates across the economy. Separate research released by Google this week found that workplace AI now touches 68 percent of jobs, representing roughly 90 percent of employment in the United States, though within any single role employees use it for only about 21 percent of their tasks on average. Scott Strand, head of strategic operations and special projects for technology and society at Google, told Axios that blue-collar workers in predominantly physical and manual roles, such as auto technicians and industrial mechanics, are increasingly relying on multimodal AI that processes images and video for real-time diagnostics and on-the-job learning.
What the Pivot Means for AI Research
The move carries implications beyond one company's org chart. AlphaFold's success was celebrated as proof that deep learning could crack previously intractable scientific problems, and its open release gave researchers worldwide free access to predicted structures for nearly every known protein. Folding that expertise into a Gemini-centric roadmap raises questions about whether future DeepMind breakthroughs will be similarly open or increasingly bundled into proprietary products.
It also signals where the talent and compute are flowing. With almost a quarter of the original AlphaFold authors gone and the remainder redirected, the institutional knowledge behind one of AI's most celebrated scientific achievements is dispersing just as the field's center of gravity shifts from specialized research models to general-purpose foundation models and agents.
The reorganization does not mean DeepMind is abandoning science. Kohli's team continues to work on enzyme design, nuclear fusion, and genomics. But the standalone AlphaFold program, the one that won a Nobel Prize, appears to be giving way to a model in which scientific discovery is meant to flow through Gemini. Whether that bet pays off, for Google and for the broader research community, will be one of the defining stories of the current AI cycle.
Stay Ahead of AI
Track the strategic moves redefining the industry with our latest AI news and analysis.
Explore more AI news
