Two more leading artificial intelligence researchers at Alphabet's Google are planning to leave for rival Anthropic, widening a talent exodus that has rattled investors and raised fresh questions about the search giant's standing in the AI race.
Jonas Adler, who worked on Google's AI coding effort, and Alexander Pritzel, who was involved in training AI systems, are both viewed internally as key contributors to the Gemini model.
The pair also worked alongside Nobel laureate John Jumper on Google DeepMind's AlphaFold research, making their exit close to the departure of an intact team.
A third researcher, Arthur Conmy, has said on X that he will join Anthropic to work on AI safety.
His LinkedIn profile lists him as a former DeepMind senior research engineer who contributed to the Gemini 2.5 model and to AI coding.
With these moves, Google has now lost four senior figures to rivals in less than a week.
A rapid succession of exits
The latest departures follow the loss of Jumper, who shared the 2024 Nobel Prize in Chemistry for AlphaFold, to Anthropic, and of Noam Shazeer, co-author of the landmark 2017 paper that introduced the transformer architecture underpinning modern AI, to OpenAI.
Shazeer's exit was a particularly bitter blow.
Google had paid more than $2 billion to bring him back in 2024 through a deal involving his startup Character.AI, only to lose him again less than two years later.
The earlier pair of departures wiped roughly $250 billion from Alphabet's market value on 22 June, the company's worst trading day in over a year.
Shares fell as much as 1.2% on Wednesday before closing only slightly lower, leaving the stock around $342.
Why the talent is leaving
The common thread running through the exits is the pull of Anthropic and OpenAI, both of which are widely expected to float in the coming months.
Joining before an initial public offering offers researchers the prospect of equity gains that are far harder to match at an established, already-listed company.
Anthropic has emerged as the chief beneficiary.
The Claude maker recently raised $65 billion at a valuation of $965 billion, overtaking OpenAI to become the world's most valuable private AI company, and has been expanding aggressively beyond general-purpose models into coding, healthcare and scientific applications.
Industry data suggests DeepMind engineers are nearly 11 times more likely to move to Anthropic than the other way round.
Money is not the only factor.
Google's top models have slipped down various benchmark leaderboards, often ranking outside the top five, while its pace of frontier releases has lagged that of Anthropic, which has launched two significant Claude Opus updates and a new class of models in the same window.
Tension over compute
The departures also come amid internal friction over access to computing power.
Shortly before Shazeer announced his move to OpenAI, computing capacity dedicated to one of his projects was reassigned to a London-based DeepMind team.
The change was said to be aimed at improving cross-team collaboration on pre-training, the early stage in which models learn from vast volumes of data.
But the allocation of scarce advanced chips has become a growing source of tension as teams compete for limited resources.
Google plays down the threat
Google has sought to project confidence.
A spokesperson said the company remained confident in its position in the market for AI talent.
DeepMind chief executive Demis Hassabis has argued that movement between leading laboratories is to be expected in the current market.
"We have by far the biggest and broadest research bench of any of the labs out there," he said at an event in Cannes.
Anthropic declined to comment.
A test of momentum
The backdrop makes the timing awkward.
Google spent much of the current AI boom playing catch-up before regaining momentum late last year through stronger Gemini models and its custom AI chips.
It is also spending heavily, having raised its 2026 capital expenditure guidance to a range of $180 billion to $190 billion and tapped markets for tens of billions in fresh equity.
For now, the practical impact of losing these specific researchers remains unclear.
But four senior defections to direct rivals in under a week is the kind of run that few investors will shrug off.
