Google has overhauled the top ranks of its artificial intelligence division in a leadership shake-up that has reordered who builds its most important technology and rattled investors. Demis Hassabis, the Nobel laureate who co-founded DeepMind, is stepping aside as chief executive of Google DeepMind, while the company's most senior scientist, Jeff Dean, is leaving after 27 years to start a new venture.
The changes, first reported on August 5, 2026 and detailed by Reuters, The New York Times, The Wall Street Journal, and The Guardian, mark the most significant restructuring of Google's AI apparatus since the company folded its Brain and DeepMind units together in 2023. Alphabet shares fell roughly 5% as the scope of the departures became clear. For a company whose future is staked on AI, the reshuffle carries weight far beyond an internal org-chart update — and AI Buzz Wire is tracking its implications across the industry.
Hassabis Moves From CEO to Chair and Chief Scientist
Hassabis will relinquish his day-to-day duties running Google DeepMind and take on two new roles: chair of DeepMind and chief scientist at its parent, Alphabet. The shift hands operational control to Koray Kavukcuoglu, a longtime DeepMind researcher who will lead the unit as senior vice president — notably, not with the CEO title.
According to The Guardian, a former Google executive framed the move as the end of DeepMind's semi-autonomous era. "The era of DeepMind as an independent actor is over," the former executive said, pointing to the appointment of a technical leader rather than a "charismatic founder-leader" as a sign that the London-based lab is being drawn firmly into the orbit of Google's Mountain View headquarters.
Hassabis co-founded DeepMind in 2010 and sold the company to Google in 2014. Under his leadership, the lab produced AlphaGo, the system that defeated the world champion of the board game Go, and AlphaFold, the protein-structure prediction tool that earned Hassabis a share of the 2024 Nobel Prize in Chemistry. In a statement reported by multiple outlets, Hassabis said he was looking forward to focusing on the "big picture" for AI and was handing over a unit well placed for the competitive battles ahead.
The reshuffle comes as Google's flagship Gemini model line has reportedly lost ground to rival systems from Anthropic and OpenAI. The same former executive told The Guardian it would be "galling" for Google that its latest Gemini iterations have slipped behind competitors after the enormous investment poured into them.
Jeff Dean's Departure After 27 Years
The leadership change was eclipsed only by the exit of Jeff Dean, Google's chief scientist and one of the most influential engineers in the company's history. Dean, who joined Google in 1999 and co-authored many of the systems that underpin modern machine learning, is leaving alongside three other senior researchers to launch a startup called Discovery Loop, according to The New York Times and TechCrunch.
The New York Times reported that four top Google AI researchers are forming the new company. Among those departing is Sanjay Ghemawat, a Google Fellow and dean of its engineering ranks who, like Dean, has been with the company for 27 years. Reporting from WIRED and Quartz described Discovery Loop as an AI-driven venture focused on accelerating scientific discovery — applying machine learning to fundamental research problems rather than to consumer products or advertising.
Dean's exit drew an emotional response inside Google. Business Insider reported that employees reacted with fond memes, and The Times of India noted that his farewell email described the decision to leave after more than two and a half decades. The Wall Street Journal characterized the departures as a "wave of exits" tied directly to the broader leadership overhaul.
Why the Shake-Up Matters
The restructuring converges on a single question: can Google keep its most talented people and stay competitive in a market where smaller, faster rivals are setting the pace?
The simultaneous departure of a founder-CEO and a generation-defining engineer is unusual even by the standards of an industry known for talent churn. Hassabis's move upward into an Alphabet-wide chief scientist role suggests Google wants his research judgment applied across the company rather than concentrated in one division. But the loss of Dean — a figure so central to Google's technical identity that he became an internal meme — underscores how the intensifying AI race is pulling talent toward independent ventures that promise more autonomy and upside.
Discovery Loop's focus on scientific discovery also signals a split in how top researchers view AI's most promising frontier. While Google and its hyperscaler peers pour resources into general-purpose chatbots and agent platforms, the researchers who built the field's foundations are betting that the next breakthroughs lie in applying AI to hard science — chemistry, biology, materials — where DeepMind itself has already proven the model with AlphaFold.
For investors, the shake-up landed at a delicate moment. Alphabet's stock decline reflected unease that the company's AI ambitions are now being pursued without some of the people most associated with its past successes. Google has since signaled continued investment, with reports of a potential large stake in an AI coding startup surfacing alongside the leadership news.
The competitive picture is unlikely to clarify quickly. Anthropic and OpenAI continue to release models that pressure Gemini's position, Meta is building a cloud business to sell AI compute, and a wave of well-funded startups is competing for the same scarce research talent. Whether Google's new structure — with Kavukcuoglu running DeepMind operationally and Hassabis advising at the Alphabet level — can close the gap will be one of the defining storylines of the coming year.
Stay Ahead of AI
Google's leadership overhaul is reshaping the competitive landscape for the world's most powerful AI labs. Follow the latest AI industry news as the fallout from this shake-up continues to unfold.
Read more AI news →
