AI & Machine Learning

Google Reshuffles AI Leadership as Demis Hassabis Steps Aside

Demis Hassabis is becoming Alphabet chief scientist as Koray Kavukcuoglu takes operational control of DeepMind and four Google veterans launch Discovery Loop.

By Olivia Grant Edited by AIstify Team Published: Updated:
Google Reshuffles AI Leadership as Demis Hassabis Steps Aside
Google is separating long-term scientific direction from DeepMind's day-to-day execution as Demis Hassabis moves to an Alphabet-wide role. Photo: Google DeepMind / Pexels

Key Notes

  • Demis Hassabis will become Alphabet chief scientist and Google DeepMind chair, leaving day-to-day operations to CTO Koray Kavukcuoglu.
  • Jeff Dean, Sanjay Ghemawat, Oriol Vinyals and Quoc Le are leaving Google to launch the machine learning laboratory Discovery Loop.
  • Google will invest in Discovery Loop and support the startup with computing capacity, preserving a strategic relationship with the departing team.

Alphabet is reorganizing the leadership of its most important artificial intelligence operation, moving Demis Hassabis out of day-to-day management at Google DeepMind and into a broader scientific role across the parent company.

Hassabis will become Alphabet’s chief scientist and chair of DeepMind, while DeepMind chief technology officer Koray Kavukcuoglu will take operational control as a senior vice president. The transition separates long-term scientific direction from the daily work of developing models, research programs and products.

The change comes alongside a consequential departure. Four senior engineers—Jeff Dean, Sanjay Ghemawat, Oriol Vinyals and Quoc Le—are leaving Google to establish Discovery Loop, a machine learning laboratory focused on automating scientific and engineering research. Google will invest in the startup and support it with computing capacity, according to Reuters.

Hassabis Moves From Operations to Alphabet-Wide Science

Hassabis co-founded DeepMind in 2010 and remained its most recognizable leader after Google acquired the laboratory in 2014. Under his leadership, DeepMind produced systems including AlphaGo and AlphaFold, then became the center of Google’s Gemini model development after merging with Google Brain in 2023.

His new position expands that mandate. As Alphabet chief scientist, Hassabis can influence research across the group’s businesses while continuing to guide DeepMind as chair. He is also expected to retain leadership of Isomorphic Labs, Alphabet’s AI drug-discovery company. The structure gives him more time for long-range research and the pursuit of artificial general intelligence while placing product execution under a separate operator.

Kavukcuoglu is not being named DeepMind CEO. Instead, the longtime research leader will oversee the organization as a Google senior vice president reporting to CEO Sundar Pichai. That distinction matters: DeepMind is a unit inside Google, and the reshuffle pulls its operating chain closer to the executive responsible for Search, Cloud and the rest of Google’s product portfolio.

Kavukcuoglu has worked at DeepMind for more than a decade and has served as CTO since 2018. His background spans both fundamental research and large-scale model engineering, making him a logical choice to connect DeepMind’s research pipeline with the commercial demands surrounding Gemini.

Discovery Loop Takes Four Google Veterans

The simultaneous creation of Discovery Loop may be more disruptive than the title changes. Dean joined Google in 1999 and helped shape the infrastructure behind products and research systems used across the company. Ghemawat was his longtime collaborator, while Vinyals and Le have led influential work in neural networks and model development.

The new public benefit company plans to apply machine learning to the research cycle itself: generating ideas, running experiments, evaluating results and repeating the process at a scale that human teams cannot easily match. That puts Discovery Loop within the expanding AI startup market for automated science rather than directly positioning it as another general-purpose chatbot developer.

Google’s investment and compute partnership soften the competitive impact. Instead of allowing four veteran researchers to move to OpenAI, Anthropic or another rival, Alphabet retains a financial relationship and potential access to the startup’s progress. It is an example of a strategy large technology companies increasingly use to keep promising teams within their wider orbit without requiring them to remain inside a corporate structure.

Google’s AI Talent Challenge Is Growing

The departures still remove substantial institutional knowledge at a sensitive point in the AI race. AIstify reported in June that Google lost Gemini co-lead Noam Shazeer to OpenAI and AlphaFold scientist John Jumper to Anthropic. Discovery Loop now extends that sequence with four people whose work reaches deep into Google’s technical history.

Elite research talent has become unusually mobile because well-funded startups can offer concentrated missions, meaningful equity and fewer organizational layers. Google can match rivals on compensation and surpass most of them in computing resources, but it must also give senior scientists enough autonomy to pursue difficult ideas. Discovery Loop suggests that even a company with Google’s research budget may need external ventures to provide that freedom.

The reshuffle therefore should not be read simply as investors losing confidence in DeepMind or Hassabis leaving Google. Hassabis is taking a wider strategic role, and Alphabet is backing the laboratory formed by the departing engineers. The more important question is whether this structure can preserve Google’s scientific ambition while improving the speed and accountability of its model and product execution.

What the Leadership Change Means

Google is effectively splitting two jobs that had accumulated under Hassabis: setting the direction of frontier science and running an organization expected to ship competitive AI products. Kavukcuoglu inherits the operating pressure around Gemini, while Hassabis gains an Alphabet-wide platform for fundamental research and AI-enabled discovery.

That division could help Google translate breakthroughs into products more consistently. It could also create coordination risk if research priorities, model schedules and business demands pull in different directions. The outcome will depend less on the new titles than on how much authority Kavukcuoglu receives and how closely DeepMind’s scientists remain connected to Google’s product teams.

Discovery Loop adds another test. If the startup produces valuable automated-research systems while maintaining a close Google partnership, Alphabet may preserve much of the upside from the departures. If it grows into an independent center of gravity for elite researchers, the move will look more like a loss of talent that Google financed on the way out.

Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.

AI & Machine Learning, News, Startups & Investment