Anthropic Poaches Four Scholars from Universities and DeepMind in Two Weeks

Anthropic announced the hiring of four leading scientists in just two weeks, between June 19 and July 1, 2026. Among them is John Jumper, the 2024 Nobel Prize winner in Chemistry and creator of AlphaFold, who left Google DeepMind after nearly nine years. It is the most high-profile episode yet in a months-long talent war between major AI labs and American universities.

Four names in two weeks makes this the most concentrated string of departures in this year's AI talent war. Jumper and his former Google colleagues, Jonas Adler and Alexander Pritzel, principal researchers on the Gemini project, moved to Anthropic nearly simultaneously, suggesting a coordinated move by an entire team rather than separate, independent decisions.

A Nobel laureate switches labs

Jumper's departure is symbolic: together with the DeepMind team, he built AlphaFold, the protein structure prediction system for which he won the 2024 Nobel Prize in Chemistry. His move to Anthropic after nearly a decade at Google DeepMind shows that even scientists at the peak of academic recognition are not immune to offers from rival labs building teams around new research directions, such as AI applications in life sciences and computational biology.

A professor on leave, not resigned

The case of Jelani Nelson, chair of the computer science department at UC Berkeley and a specialist in streaming algorithms and dimensionality reduction, looks different. Nelson joined Anthropic on a research leave, keeping his university position, which in theory leaves him a path back. In practice, however, arrangements like this give AI labs access to academic networks and student talent without fully severing ties with the university.

Universities worried about brain drain

Since the start of 2026, at least 22 professors and researchers have taken leave or left their universities to join top AI labs, according to figures cited by The Information. Faculty who remain warn that the drain of talent threatens the development of open AI models in the West, since universities are responsible for much of the foundational research that commercial systems later build on.

This isn't an ordinary vendor-customer relationship, it's co-designing the memory architecture that future Claude models will run on - Anthropic, on its partnership with Micron, cited in the context of the company's research infrastructure strategy

The bigger picture of the talent race

Anthropic's case is not an isolated one. Google DeepMind has lost six researchers from its coding team in recent months to Meta, OpenAI, and Anthropic, while Noam Shazeer, co-author of the landmark Transformer architecture paper and co-lead of the Gemini project, had earlier left Google for OpenAI. Meta has built its own Superintelligence Labs, offering compensation packages worth tens of millions of dollars to lure top researchers from competitors.

For Poland's scientific and academic community, the signal is clear: competition for the best AI researchers is shifting from raw model capability toward algorithmic theory and research infrastructure, areas where specialized university scientists are becoming especially sought after by commercial labs. Polish universities and institutes, including those involved in developing domestic language models, will have to contend with similar recruitment pressure from global companies, though for now on a far smaller financial scale.

Sources: Universities Fret as Anthropic, OpenAI, Meta and DeepMind Lure Their Professors (theinformation.com), Anthropic Poaches Four Top Scholars in Two Weeks (finance.biggo.com), Google DeepMind's Coding Pivot Lost Six Researchers to Meta, OpenAI, and Anthropic (techtimes.com)