Two senior researchers closely associated with Google DeepMind’s Gemini program are preparing to join rival AI lab Anthropic, according to Bloomberg reporting, citing people familiar with the matter.
The individuals, Jonas Adler and Alexander Pritzel, are both considered significant contributors within DeepMind’s model development efforts. Adler has worked on AI coding systems, while Pritzel has focused on pretraining, the foundational stage where large models learn from vast datasets to develop general capabilities.
A separate departure adds further weight to the trend. John Jumper, who shared the 2024 Nobel Prize in Chemistry for AlphaFold, is also reported to be leaving Google DeepMind after nearly nine years to join Anthropic, underscoring the scale of talent movement at the top of the field.
Taken together, the exits reflect a broader pattern of mobility across frontier AI labs, where leading researchers increasingly shift between organizations shaping the next generation of foundation models.
High-Profile Exits Deepen Scrutiny of DeepMind’s Talent Retention
The reported departures come amid a series of notable exits from Google’s AI divisions, particularly within teams responsible for landmark machine learning breakthroughs.
Among the most prominent was Noam Shazeer, a senior engineering figure and co-author of the influential “Attention Is All You Need” paper, who left to join OpenAI. His exit was widely interpreted as symbolically significant, given his role in developing the transformer architecture that underpins most modern large language models.
Jumper’s move has drawn even greater attention due to his scientific credentials and impact. As a key contributor to AlphaFold, his work helped solve one of biology’s long-standing challenges: protein structure prediction. His departure, following years of recognition at DeepMind, has intensified discussion around retention at the organization.
The broader impact has also been felt in financial markets. Alphabet Inc. experienced a sharp decline in market value following news of senior exits, reflecting investor sensitivity to perceived shifts in leadership and research momentum within its AI division.
While Alphabet Inc. continues to maintain one of the most advanced and well-funded AI infrastructures globally, repeated high-profile departures are increasingly being viewed as part of a more structural pattern rather than isolated incidents.
AI Competition Tightens as Anthropic and Rivals Accelerate Hiring
The movement of researchers between leading AI labs reflects a broader intensification of competition across the sector. What was once seen as a relatively clear lead held by OpenAI has evolved into a far more contested and dynamic landscape.
Major technology companies, including Microsoft and Google, have invested heavily in AI infrastructure, model training, and recruitment, spending billions to secure long-term positioning in foundation model development. These investments extend beyond compute resources into aggressive hiring strategies aimed at attracting researchers at the frontier of capability development.
Anthropic, meanwhile, has positioned itself as a focused challenger in areas such as coding systems, scientific reasoning, and AI safety. Its hiring strategy has increasingly targeted researchers with expertise in pretraining and system-level optimization, capabilities seen as critical for improving model efficiency and performance.
Financial incentives remain an important factor, but they are no longer the sole driver of mobility. With both Anthropic and OpenAI moving toward potential public listings, equity upside has become a growing consideration for senior researchers evaluating long-term opportunities.
Access to compute is also emerging as a defining constraint. As demand for specialized hardware such as tensor processing units increases, competition over infrastructure is shaping where researchers can most effectively conduct frontier work.
As one Chandra Goud, Consultant at The Perfume Shop, observed: “John Jumper leaving DeepMind for Anthropic is the biggest talent move in AI since Noam Shazeer went back to Google. Nobel laureate. The man who solved protein folding. Gone.”
Goud went on to describe just how significant these moves could be for the AI race:
“Within 12 months, Anthropic will announce a science-focused model that isn’t just chat — it’s a research agent that designs and iterates on experiments autonomously. DeepMind’s brain drain is now undeniable.”
The comment reflects growing speculation that talent flows are not just about compensation or prestige, but about differing visions for how frontier AI systems should be applied.
Talent Mobility Signals a Shifting Center of Gravity in AI
While Google DeepMind retains significant institutional strength, continued losses of high-profile researchers risk shaping external perceptions about momentum and direction. In a field where progress is closely tied to individual expertise, even a small number of exits can carry disproportionate signaling effects.
Anthropic’s sustained recruitment from rival organizations suggests it is increasingly positioning itself as a destination for people wanting to be at the forefront of AI. That positioning is likely to remain attractive as frontier AI development continues to demand greater compute, talent, and specialized expertise.
Looking ahead, the key question is not whether individual researchers will continue to move between companies, but how sustainable current organizational models are in retaining top-tier talent. The outcome may help determine which labs lead the next phase of AI development.
As competition intensifies and the stakes grow, the movement of researchers between DeepMind, OpenAI, and Anthropic is becoming one of the clearest indicators of where momentum in the AI sector is heading next.
