Google DeepMind researchers defect to Anthropic

The Adler and Pritzel news lands as the latest beat in a strikingly compressed sequence. In roughly seven days in June 2026, Google lost at least four high-profile AI staffers to rivals ^[5]. The chain reads almost like a countdown: on June 19, Noam Shazeer, a Gemini co-lead and co-author of the 2017 'Attention Is All You Need' paper, announced he was leaving for OpenAI ^[5]. On June 21, Nobel laureate John Jumper said he would join Anthropic ^[6]. Then on June 24, Bloomberg reported that Adler and Pritzel, both internally regarded as key Gemini contributors, were also Anthropic-bound ^[1]. What makes the cluster sting is the seniority and the specificity of the losses. Adler led work on Google's AI coding effort and Pritzel worked on pretraining, the foundational stage in which models learn from large volumes of data ^[5]. These are not peripheral hires but people sitting close to the core of how Gemini gets built. TechCrunch framed the run plainly: AI researchers continue to leave Google for its rivals ^[2].

The most interesting disagreement in the coverage is over motive. The obvious explanation is financial: Anthropic and OpenAI are edging toward potential public listings, dangling rare pre-IPO equity that established Big Tech salaries struggle to match, and Anthropic has reportedly raised $65B at a $965B valuation ^[1][5]. But Fortune's AI editor Jeremy Kahn pushes back on the simple cash story, arguing that money is unlikely the driver, especially for Jumper, and that the moves instead signal 'the scientific opportunity at Anthropic was actually better-which is much worse news for Google DeepMind' ^[3]. There is also a friction-from-within thread: cryptobriefing reports that shortly before Shazeer's move, Google reassigned computing capacity from one of his projects to another DeepMind team in London, with resource allocation becoming a recurring tension as teams compete for limited advanced chips ^[5]. Read together, the picture is less a bidding war and more a pull toward nimbler organizations where senior scientists feel they can move faster and access the resources they want.

Wall Street did not treat the departures as noise. Alphabet shares fell roughly 5-7% after the Shazeer and Jumper announcements, with cryptobriefing pegging the wipeout at around $270B in market value ^[3][5]. The deeper anxiety is competitive: industry watchers question whether DeepMind can stay in the lead pack, noting its top models are reportedly ranked outside the top five on various benchmark leaderboards, behind Anthropic and OpenAI ^[3]. D.A. Davidson analyst Gil Luria put it bluntly, saying the race at the frontier now appears to be between Anthropic and OpenAI ^[4]. The counterweight comes from inside and from some of the Street. Demis Hassabis insists Google has 'by far the biggest and broadest research bench of any of the labs out there' and continues to win its fair share of talent ^[4], while Jefferies analyst Brent Thill dismissed the wave as 'an industry-wide war for talent' and kept a Buy rating on Alphabet ^[3]. The gap between the stock reaction and the executive reassurance is, for now, the live question.

The social conversation complicates the tidy 'Google is collapsing' narrative. On X, Bloomberg's Ed Ludlow and AI aggregator accounts amplified the Adler and Pritzel report and tied it to the Jumper and Shazeer wave, and search even surfaced a separate Gemini 2.5 contributor announcing a move to Anthropic for safety work. But on Reddit, the framing splintered. One ML engineer argued the net migration of talent may actually run from OpenAI toward Google, and that Anthropic stands out simply because it is such a clear winner that it attracts the most attention, not because Google is uniformly losing. A second recurring theme was culture over cash: commenters disputed the idea that Anthropic is outbidding the giants and pushed a mission-and-freedom thesis, captured by a widely shared line that you cannot buy talent or culture. The unconfirmed claim that DeepMind is broadly struggling, sourced to Twitter speculation, was flagged by posters themselves as unverified. The takeaway is that the headline exits are real and consequential, but the 'who is winning the war for talent' verdict is far less settled than the stock chart suggests.

The June 2026 window in figures: at least four high-profile Google AI departures in roughly one week, with Shazeer to OpenAI and Jumper, Adler, and Pritzel to Anthropic ^[2][5]. Alphabet's stock fell roughly 5-7% on the news, a swing cryptobriefing sizes at around $270B in market value ^[3][5]. The recruiting backdrop: Anthropic reportedly raised $65B at a $965B valuation, while Shazeer's own 2024 return to Google came via a roughly $2.7B Character.AI licensing and acqui-hire deal ^[2][5]. The throughline is scale: a handful of named exits, a quarter-trillion-dollar market reaction, and a pre-IPO equity lever that Big Tech compensation cannot easily replicate.