Within a span of a few days Google lost two prominent researchers: Noam Shazeer, a co-author of Transformer whom DeepMind staff credit with rescuing Gemini, left for OpenAI; and John Jumper, described in the report as a Nobel laureate and the leader of AlphaFold, joined Anthropic. The rapid succession of exits has focused attention on internal issues at Google that may have played a larger role than external offers.
Why this matters
Google possesses deep financial resources, a broad product and infrastructure stack, and more distinguished awardees than many rivals, yet it saw two highly regarded figures depart in a short period. Multiple reports and observers argue that market-driven pay competition alone does not fully explain these departures.
Indications pointing inward
Reports highlight two main internal factors:
-
Large-company sclerosis and internal competition: turf wars over TPU (Tensor Processing Unit) capacity among merged teams, parallel development efforts — including three groups shipping the same coding agent — and promotion practices that reward launching new initiatives over sustaining existing projects.
-
Leadership decisions and partial commitment: according to accounts, company leadership appeared not fully convinced about its AI bet; one cited example is that last October Google rented prized TPU capacity to Anthropic instead of prioritizing its own DeepMind.
Together these factors suggest that the pull of pre-IPO equity and richer offers is only part of the story; organizational frustration and structural barriers are important push factors as well.
Conclusion
These departures illustrate that even a company with vast spending power and technological assets can lose top talent if its internal organization and leadership do not effectively support core teams. Observers interpret the exits of Noam Shazeer and John Jumper through the lens of Google’s internal structures and decisions: the leak may be organizational dysfunction rather than compensation alone.
