Less than a year after AlphaFold’s results brought a Nobel Prize to Google DeepMind, the company has broken up the original research team. Over the past year most of the authors of the landmark paper were reassigned to other roles, and nearly a quarter of the original group have left the company. The AlphaFold work has been partly folded into a broader push around Gemini.
What changed
- During an internal reorganization, the majority of the researchers credited on the AlphaFold paper were moved into new positions or reassigned to different projects over the past year.
- Nearly 25 percent of the original contributors have departed DeepMind in that period.
- Pushmeet Kohli, DeepMind’s Vice President of Research, told the Financial Times that the lab’s nine-year strategy of targeting major scientific challenges "has evolved."
- Nobel laureate John Jumper left DeepMind for Anthropic in June, followed by two other core AlphaFold researchers.
Why this matters
AlphaFold produced a major scientific breakthrough by addressing a decades-old problem in protein structure prediction, and its methods are now widely used in drug discovery. Despite that, DeepMind chose not to keep the original team intact and instead integrated the work into a wider Gemini-focused strategy.
The shift is notable because it shows that historic scientific success — even one recognized at the highest level — does not necessarily preserve a dedicated research team within a corporate AI lab when strategic priorities change.
How the strategy is being rewritten
According to company statements, DeepMind is reallocating people and resources toward Gemini, a program focused on large-scale language and multimodal models. Leadership sees greater value in concentrating expertise around a broad AI platform than in maintaining a single, problem-specific research group.
Consequences
As a direct result, some AlphaFold researchers have moved to other organizations or been absorbed into Gemini projects. Over time, this could reshape how industry-funded teams organize to tackle large scientific challenges: the pace and direction of corporate AI competition may determine which research directions receive sustained, dedicated support.
Conclusion
DeepMind’s decision indicates that corporate strategy can override past scientific achievements. A Nobel Prize and a proven technical accomplishment do not guarantee that a team or its mission will remain unchanged when a company pivots its priorities.
