Google DeepMind loses Nobel laureate John Jumper to Anthropic, its third senior researcher to depart in as many months

Google DeepMind Vice President John Jumper, who shared the 2024 Nobel Prize in Chemistry with CEO Demis Hassabis for developing AlphaFold, announced on Friday that he is leaving the company after nine years to join Anthropic, becoming the third senior DeepMind researcher to leave in as many months, and that his recent work at Google was spent on artificial intelligence coding tools rather than the scientific work that earned him the Nobel.

Nine years later: An unexpected breakout for John Jumper

John Jumper confirmed the move in a post on X, framing it as the end of a long chapter rather than a clean break. “After nearly 9 years, I’ve decided to leave Google DeepMind and join Anthropic (after taking some time to recharge). I’m incredibly grateful for my time at GDM. Demis Hassabis took a real chance by letting me lead the AlphaFold team just six months after finishing my PhD, and the entire GDM team taught me so much about how to do great science. GDM is a special place, and I’ll still be excited to hear what amazing things they’ll discover next.”

John Jumper was assigned to the AlphaFold team barely six months after completing his PhD and went on to lead the system that can predict the three-dimensional structure of a protein directly from its amino acid sequence; It’s a tool now credited with predicting more than 200 million protein structures, disrupting biological research for months and sometimes years.

From science to software: a quiet transition

According to a report, John Jumper stepped away from pioneering science work before leaving and moved to Google’s AI coding development team. Bloomberg The report comes as Google tries to close the gap with Anthropic, OpenAI and other rivals when it comes to selling AI coding tools to businesses.

The change takes the Nobel laureate’s latest Google post slightly away from the protein research that made his name and underscores how tenuous Google’s research priorities have become as the company struggles on multiple fronts at once.

Farewell note: gratitude, not complaints

Hassabis responded publicly and credited AlphaFold with reshaping expectations of what AI could do for science.

“What we achieved with AlphaFold changed the world and showed the field what is possible with AI for science and medicine, illuminating the way for how AI can benefit humanity,” he wrote, a tribute that reflected the warmth of Jumper’s own post rather than any harsh statement.

Three outs, one quarter: loss of talent

Jumper’s move is the third high-profile departure from Google DeepMind in as many months. Days earlier, Noam Shazeer, co-leader of Google’s Gemini model and co-author of the 2017 paper introducing the transformer architecture that underpins most modern AI systems, had announced he was leaving for OpenAI, less than two years after Google paid a reported $2.7 billion to bring him back from Character.AI.

“I am excited to share that I will be joining OpenAI and look forward to working with the incredible team there,” Shazeer wrote, adding that “it was a difficult decision to move on” and that he was “incredibly proud of the amazing team at Google.”

Before that, David Silver, the principal researcher behind AlphaGo, AlphaZero, and MuZero, left DeepMind to found his own startup, Ineffable Intelligence, which pursues reinforcement learning approaches to superintelligence rather than large language models.

Anthropic’s hiring spree: Karpathy also joins

The hiring of John Jumper extends a number of significant additions to Anthropic into 2026. In May, Andrej Karpathy, co-founder of OpenAI and former head of AI at Tesla, joined the company from his startup Eureka Labs to form a team focused on using Claude to accelerate its own pre-education research.

“I think the next few years will be particularly formative at the graduate level. I am very excited to join the team here and return to R&D,” Karpathy wrote at the time.

John Jumper didn’t say what his role at Anthropic would be, only that he planned to take time off first. The moves come as both Anthropic and OpenAI are gearing up for possible public listings, intensifying competition among Silicon Valley’s biggest AI labs for the small pool of researchers who can work at the frontier.

AGI timeline: ‘a few years away’

The departure comes just days after Hassabis, DeepMind’s chief executive, used a fireside chat at the Stanford Graduate School of Business published online Tuesday to argue that humanity has little time to prepare for the arrival of artificial general intelligence, defined as artificial intelligence that can match or exceed human performance on cognitive tasks.

Demis Hassabis gave a rough date for this turning point. “Maybe 2030, plus or minus a year, it’s really surprising to think about. I think it will be a huge transformative technology; it will effectively be a new era of humanity,” he said, likening its arrival to a singularity, a threshold of no return.

New human age: preparation for AGI

Demis Hassabis, whose comments placed him among a group of leading AI executives including OpenAI’s Sam Altman and Anthropic’s Dario Amodei who warned that society has a narrow window to adapt, said some of his industry colleagues had become “very confident” in their predictions. Still, he noted AGI’s potential to lead to medical breakthroughs and economic transformation and raise the possibility of a “post-famine world.”

He urged students in both humanities and STEM fields to “lean in” to technology rather than resist it. “The community needs to hear this because we don’t have a lot of time to prepare for what this means,” he said.

Demis Hassabis closed by framing the coming years as a real crossroads rather than a fixed outcome. “The future is still to be written in my view, but the next few years will be very critical in terms of which direction this goes and what we collectively want this to look like,” he said.