In Brief
Google announced that Demis Hassabis is stepping out of the CEO role at Google DeepMind, its AI research lab, the same week chief scientist Jeff Dean left after 27 years to found a startup. Alphabet stock closed down about 4%, capping an eight-month exodus of Google's most important AI talent.
What Happened
Google CEO Sundar Pichai announced that Demis Hassabis is stepping out of the CEO role at Google DeepMind, the company's AI research lab and the group behind its Gemini models. Hassabis becomes chair of Google DeepMind and chief scientist of Alphabet, Google's parent company, while staying on as CEO of Isomorphic Labs, the drug discovery company that spun out of DeepMind. Hassabis wrote that handing off day-to-day operations frees him to focus on the big picture.
Koray Kavukcuoglu, previously Google DeepMind's chief technology officer, takes over as senior vice president of Google DeepMind. He reports directly to Pichai and now oversees Gemini model development, frontier AI research (the work of building the most capable models anywhere), and the Gemini app and developer teams.
Separately, Jeff Dean, chief scientist across Google DeepMind and Google Research, announced his exit after 27 years. Dean helped create Google's first ad system, launched Google News and Translate, started the TPU chip program (Google's custom AI hardware), and co-founded Google Brain, the company's early AI research team. He is founding a startup called Discovery Loop with a group of fellow ex-Googlers to automate machine learning, science, and engineering. Alphabet stock closed down about 4% the day the news broke.
SmarterX founder and CEO Paul Roetzer explained why the shakeup was months in the making, and what it says about where power now sits at Google, on Episode 230 of The Artificial Intelligence Show.
The Key Numbers
27 - Years of Jeff Dean's tenure at Google before founding Discovery Loop
4% - Alphabet's stock drop the day the news broke of Dean's departure and Hassabis's new role
8 - Span, in months, of the talent exodus from Google DeepMind
$2.7 billion - Google's price to bring back Noam Shazeer, who later left for OpenAI
2030 - Hassabis's earliest time estimate for the arrival of AGI
Why Power Is Shifting From London to Silicon Valley
The exodus started eight months ago. Hassabis and Dean are not the first to step back from or leave their positions. In February, longtime DeepMind researcher David Silver, a key builder of AlphaGo (DeepMind's AI that beat the world's best Go players), left to found his own startup.
In June, Noam Shazeer, lead author of the Transformer paper (the 2017 Google research breakthrough that underpins ChatGPT and every modern chatbot), left for OpenAI. John Jumper, who shared a Nobel Prize with Hassabis for AlphaFold, DeepMind's protein-structure AI, left too.
The replacement got an SVP title, not CEO. "One thing that jumped out to me right away on the Hassabis change is that their replacement is a senior vice president," says Roetzer. "So they are not putting a new CEO in place, which tells you how DeepMind kind of fits in the overall structure here."
DeepMind is headquartered in London. The Financial Times reported that "Google is shifting control of its AI effort from London back to Silicon Valley," describing "a seismic shift" as "commercial urgency eclipses the research-led culture that defined DeepMind."
Roetzer saw it coming in June. On Episode 221, after reading the book The Infinity Machine, Roetzer said: "There's so many times when reading the book that I found myself thinking Demis might leave at some point if he no longer felt that Google was the best place to pursue his research and mission."
The signals kept coming. Hassabis closed Google's I/O conference in May talking about being at the foothills of the singularity, the point where AI progress outruns human comprehension. He later told Axios he believes AGI, which is AI about as intelligent as humans, could arrive as soon as 2030 and AI proving to be 100 times as impactful as the Industrial Revolution.
Sergey Brin is stepping back in. At a Stanford panel last fall, Brin, the Google co-founder admitted, "I guess I would say in some ways we messed up in that we under-invested and didn't take it as seriously as we should have maybe eight years ago when we published the transformer paper." Now, per the Financial Times, Brin has re-emerged as one of the company's most influential voices on AI strategy.
Roetzer points back to December 2025, when OpenAI declared an internal code red because Google's Gemini 3 had surpassed its models on industry benchmarks. Eight months later, Google's models have fallen off the top of the charts and its leadership is scattering.
"It's a very, very interesting time, and eight months is like eight years in AI time because it has dramatically changed from what we were saying in December '25."
— Paul Roetzer, founder and CEO of SmarterX, Episode 230 of The Artificial Intelligence Show
SmarterX Take
Roetzer believes the change is good for society. Hassabis, a researcher at heart, gets to focus on AGI, scientific discovery, and curing disease at Isomorphic Labs.
The bigger question is whether the frontier race even matters for Google. "Do they care to be on the frontier? Do they need to have the most powerful model versus Claude or ChatGPT? They have the distribution," Roetzer says. For businesses that touch Gemini mainly through Google Workspace, the model may only need to be good enough.
SmarterX lives that reality as a Google enterprise customer, with Gemini embedded in Drive, Gmail, Docs, and Sheets every day. But for AI agent capabilities and advanced reasoning models, the kind that work through complex multi-step problems, "nine times out of 10 we're going to Claude Code or ChatGPT to do that kind of work," Roetzer says. Distribution keeps Gemini in the workflow, but it has not yet made it the model of choice for the hardest work.
What to Watch
Google's next model release will test the reorganization. Google is selling the reorg to markets as part of the plan: deep bench, accelerated research, everything on track. Roetzer says that it might be, if strengths in infrastructure, data, and distribution produce a real leap in model capabilities. If not, the questions about the exodus intensify.
Compute allocation inside Alphabet is the next fight. Roetzer says internal competition for compute, which is the processing power that trains and runs AI models, is a defining challenge. How much goes to Isomorphic Labs, to research, and to safety and alignment work when the same infrastructure powers Search, Ads, and products serving billions of people? Those decisions will shape what Google's AI can do next.
Which AI Platforms Are Companies Using Most?
Google Gemini is provided by 42% of organizations, trailing OpenAI's ChatGPT at 59% and Microsoft Copilot at 46%, according to the 2026 State of AI for Business Report. Gemini is strongest at the smallest firms, reaching 62% at companies under $1 million in revenue, then falls to the low 20s at $500 million and up. Distribution gets Google in the door, but the platform war is far from settled.
The full report, based on 2,100+ responses from professionals across roles, functions, and industries, covers where organizations stand on adoption, governance, training, and the tools they provide. Read the full report →
Mike Kaput
Mike Kaput is the Chief Content Officer at SmarterX and a leading voice on the application of AI in business. He is the co-author of Marketing Artificial Intelligence and co-host of The Artificial Intelligence Show podcast.