AI Talent Drain: Google Researchers Join Anthropic

What happened?

According to TechCrunch, AI researchers Jonas Adler and Alexander Pritzel have left Google to join Anthropic, the rival startup backed by Amazon. This departure adds to a series of high-profile exits, including those of Noam Shazeer (co-author of the 'Attention Is All You Need' paper) and John Jumper (AlphaFold lead), who also joined Anthropic in previous months. Adler and Pritzel were key figures in Google DeepMind teams, working on foundation models and reinforcement learning. Their departures, though not as high-profile as Shazeer's or Jumper's, reflect a consistent trend: Google is losing some of its best brains in artificial intelligence.

Why is it important?

Google's talent drain is not a new phenomenon, but it has intensified over the past year. The company has lost several of its most prominent researchers, lured by startups like Anthropic that offer greater autonomy, a pure research culture, and competitive compensation packages. This exodus reflects a broader industry trend: AI companies are struggling to retain experts in an extremely tight labor market, where annual salaries can exceed one million dollars. According to data from Levels.fyi, senior AI researchers at startups like Anthropic can earn between $800,000 and $1.5 million annually in total compensation, while at Google packages are often lower due to corporate structure. Additionally, Google's research culture has been affected by bureaucracy and pressure to commercialize products, contrasting with Anthropic's academic focus that prioritizes safety and fundamental research.

Historically, Google has been a hotbed of AI innovation, with milestones like the 'Attention Is All You Need' paper (2017) that gave rise to transformers, the basis for models like GPT and Gemini. However, since 2022, the company has seen a steady outflow of talent to startups and competitors. For example, in 2023, Google Brain co-founder Andrew Ng had already left the company years earlier, and in 2024, several DeepMind researchers joined startups like Mistral AI and Cohere. This drain not only affects Google but also strengthens its rivals, creating a more fragmented ecosystem.

Consequences for Google and the industry

For Google, the loss of talent could erode its competitive advantage in AI, especially in areas like language models and quantum computing. Although the company still has enviable depth of talent, the departure of key figures could slow the development of products like Gemini and its competitors. According to an analysis by consulting firm Gartner, turnover of senior researchers can delay critical projects by 6 to 12 months, as tacit knowledge is lost. Additionally, the morale of remaining teams may be affected, creating a domino effect. For Anthropic, these hires strengthen its position as a serious contender, accelerating its research in safe and aligned AI. The startup, which raised $7.5 billion from Amazon in 2025, now boasts an elite team that includes former researchers from Google, OpenAI, and DeepMind. This allows it to compete directly with OpenAI and Google in developing frontier models.

At the industry level, the war for talent could further inflate salaries and increase turnover, complicating long-term planning. Big tech companies are responding with strategies such as creating more autonomous internal labs (e.g., Google DeepMind already operates with some independence) and offering record compensation packages. However, startups have the advantage of offering equity in companies that could become hugely valuable, like Anthropic, whose valuation is estimated at $60 billion in 2026. This dynamic echoes the Web 2.0 talent bubble in 2010, when engineers from Google and Facebook left for startups like Uber and Airbnb, transforming the tech landscape.

What should readers know?

Readers should understand that this drain is not a failure of Google, but a symptom of a hypercompetitive market. Google remains a leader in research publications and has massive resources, but corporate culture and bureaucracy can discourage researchers seeking immediate impact. Moreover, the trend of researchers moving to startups suggests that the AI ecosystem is decentralizing, with new companies emerging as innovation hubs. For users, this could translate into more diverse and safer products, as Anthropic focuses on aligned AI, while Google prioritizes integration with its existing services. In the job market, AI professionals should watch for opportunities at startups, which offer greater autonomy and growth potential, albeit with higher risk. In summary, the departure of Adler and Pritzel is a reminder that talent is the scarcest resource in AI, and its mobility is redefining the balance of power in the industry.