Google AI researchers Jonas Adler and Alexander Pritzel leave for Anthropic

Google is set to lose two more senior artificial intelligence researchers to Anthropic, Jonas Adler and Alexander Pritzel, in another blow to the team behind Gemini. Their planned exits deepen a streak of departures that is testing how long Google can keep elite model builders inside one shop while rivals use money, mission and speed to pry them away.

Bloomberg identified Adler and Pritzel as key contributors to Gemini. Adler worked on Google’s AI coding efforts, while Pritzel was involved in training AI systems, two parts of the stack that matter when frontier models are being tuned for both capability and reliability. Their moves follow Noam Shazeer’s departure from Google to OpenAI and John Jumper’s move from Google DeepMind to Anthropic after nearly nine years at Google, leaving Google to defend not only product momentum but the bench of people who know how that product was built. Jumper, who shared the 2024 Nobel Prize in Chemistry for AlphaFold, gave Google one of its most visible scientific credentials in AI.

The departures land as Google continues to push Gemini 3.5. Google highlighted the model at I/O 2026 in May, presenting it as part of a broader push into agentic tools and faster model deployment. That makes the personnel churn more than a human-resources story: if the company cannot hold on to the researchers closest to Gemini’s coding and training systems, it risks slowing the cadence of new releases and weakening its leverage in a talent market where senior engineers can move quickly to better-funded rivals.

Anthropic is also sending a clear signal about why it can recruit. In April, the company said it had signed a new agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity expected online starting in 2027. Anthropic also said a separate expansion of its Google Cloud TPU use was worth tens of billions of dollars and would bring well over a gigawatt online in 2026. The company describes itself as an AI safety and research company building reliable, interpretable and steerable systems, a positioning that may appeal to researchers who want a narrower mission than Google’s sprawling product empire. The immediate question for Google is whether compensation alone can counter that pull, or whether bureaucracy, mission and execution speed are proving just as important.