Who Is Demis Hassabis, Chair of Google DeepMind?
The Nobel laureate, former DeepMind CEO, and Alphabet chief scientist leading Isomorphic Labs.
Demis Hassabis is the chair of Google DeepMind, chief scientist of Alphabet, and founder and chief executive officer (CEO) of Isomorphic Labs.
At Layer3Labs, we build artificial intelligence (AI) agents for business workflows, so we follow leadership changes at the labs that build frontier models.
On August 5, 2026, Hassabis handed day-to-day management of Google DeepMind, which he had led since co-founding DeepMind in 2010, to Koray Kavukcuoglu. He now works with Sundar Pichai on artificial general intelligence (AGI) strategy and is leaning into his role at Isomorphic Labs.
Where Demis Hassabis Is Now
Demis Hassabis no longer runs the day-to-day operations of Google DeepMind. On August 5, 2026, Google announced that Koray Kavukcuoglu, now senior vice president (SVP) of Google DeepMind reporting to Sundar Pichai, would take over day-to-day running. Hassabis became Chair of Google DeepMind and Chief Scientist of Alphabet, and he continues to lead Isomorphic Labs as CEO.
Fortune reported that Alphabet shares fell around 5% following the announcement. Hassabis took the Alphabet chief scientist title as Jeff Dean left Google to co-found Discovery Loop, a transition detailed in our profile on Jeff Dean.
- Current roles: Chair of Google DeepMind, Chief Scientist of Alphabet, and Founder and CEO of Isomorphic Labs.
- Operating company: Alphabet Inc.
- Known for: Co-founding DeepMind, leading DeepMind through AlphaGo and AlphaFold, and receiving the 2024 Nobel Prize in Chemistry.
- Education: Double first in the Computer Science Tripos at Queens' College, Cambridge (1997); PhD in cognitive neuroscience from University College London (2009).
- In the news: Handing operational control of Google DeepMind to Koray Kavukcuoglu on August 5, 2026.
November 2026 Remarks on AI Oversight
Demis Hassabis said an incident he sometimes calls an oil spill is going to happen on the current trajectory. He made that warning conditional on the industry collectively paying a lot more attention to monitoring and robustness. Masters of Scale says the episode was published November 6, 2026. In its introduction to the conversation with Reid Hoffman, the podcast identifies Hassabis as the co-founder and chair of Google DeepMind and chief scientist of Alphabet.
The warning followed a discussion of cyber issues and AI agents. Hassabis said, "We're also building much more hardened sandboxes. I think a lot of the cyber issues we've seen in the last six months or so could have been addressed with better engineering, more robust engineering, and better monitoring systems." He added, "It shouldn't be possible for agents to be running for weeks without anyone realizing it."
Hassabis then said, "Some kind of incident, I sometimes call it like an oil spill, is going to happen on the current trajectory if we don't collectively pay a lot more attention and put a lot more resources into these types of systems." He gave no date for such an incident and named no incident.
The October 2026 warning on AI agent hacks came from a different outlet's report of a separate statement.
Hoffman also asked about recursive self-improvement (RSI), a process in which AI helps improve AI systems. Hassabis said the technique mostly works best in verifiable domains such as math and coding. For now, he said, human engineers remain involved in reviewing code, designing systems and writing code.
Hassabis drew a distinction between that work and closed-loop RSI without human review. He called the closed-loop approach "a very dangerous way to get to full AGI." He said researchers should be cautious as they approach closed-loop RSI because control of the process and understanding its end result could be difficult.
If you run AI in your own workflows, use the human-in-the-loop AI accuracy guide to review where people should check its work.
Hassabis's proposed safeguards covered engineering, testing and oversight:
- Build more hardened sandboxes and improve engineering and monitoring.
- Eventually, seek formal guarantees, maybe bounds on what systems can do, and tests of guardrail robustness.
- Consider an industry-run standards body with government oversight, which Hassabis said he thinks will be needed, while recognizing that geopolitics makes an international, U.S.-led body difficult.
Executive Reorganization at Google DeepMind
The leadership reorganization announced on August 5, 2026, separated day-to-day running of Google DeepMind from Hassabis's new AGI strategy role. In a message to staff, Hassabis stated that "now is the right time for me to hand over my day-to-day operational responsibilities at GDM." Sundar Pichai described the move as "a role that allows him to put his full attention on actively shaping the future of AGI."
Koray Kavukcuoglu, previously Google DeepMind's chief technology officer (CTO) and Google's chief AI architect, became Senior Vice President (SVP) of Google DeepMind reporting to Sundar Pichai. As reported by Fortune, Kavukcuoglu does not hold a stand-alone CEO title. Kavukcuoglu oversees Gemini model development, frontier AI research, and the Gemini application and developer teams.
In his farewell note as CEO, Hassabis cited the Gemini application reaching over 950 million monthly users and named Gemini 4 among upcoming models. Teams evaluating these systems can consult our guides on how to use Gemini and our Gemini 3.1 Pro review.
October 2026 Warning on AI Agent Hacks
The News International reported on Wednesday, October 14, 2026, that Demis Hassabis said recent hacking incidents carried out by artificial intelligence (AI) agents are early warning shots of an automated era. The report is by Ruqia Shahid. The outlet quoted the Google DeepMind co-founder and chair directly: "I think these (hacking incidents) are early warning shots for us that systems aren't that powerful yet, so the incidents aren't that major."
Hassabis said the incidents could become major over the next year and the next couple of years if industry, individual companies, and governments do not do more work. He pointed to cross-pollinating ideas from fields like biology or the nuclear sector to secure critical systems, and noted that an international regulatory framework would eventually be required. The News International did not name the conference where he spoke or itemize the hacking incidents.
Video Game Engineering and Design
Before entering academic neuroscience or artificial intelligence research, Demis Hassabis worked as a video game programmer and studio founder. According to his biographical record on Wikipedia, he achieved chess master standard as a child with an Elo rating of 2300, and later won the Pentamind world championship at the Mind Sports Olympiad five times between 1998 and 2003.
As a teenager, Hassabis co-designed and lead-programmed the 1994 game Theme Park with Peter Molyneux at Bullfrog Productions. Theme Park sold several million copies.
Hassabis completed the Computer Science Tripos at Queens' College, Cambridge, graduating in 1997 with a double first. He then worked as lead AI programmer on Black & White at Lionhead Studios and founded the London games developer Elixir Studios in 1998. Hassabis served as executive designer on Republic: The Revolution and Evil Genius until Elixir Studios closed in 2005.
Cognitive Neuroscience and Memory Research
Demis Hassabis then moved into cognitive neuroscience, where his research focused on memory and imagination. He earned a PhD in cognitive neuroscience from University College London (UCL) Queen Square Institute of Neurology in 2009 under the supervision of Eleanor Maguire.
His research showed that patients with damage to the hippocampus could not imagine new experiences. The work linked memory and imagination, and the journal Science listed it among the top 10 scientific breakthroughs of 2007.
Hassabis completed postdoctoral research at the Massachusetts Institute of Technology (MIT) and Harvard University before securing a Henry Wellcome fellowship at the UCL Gatsby Computational Neuroscience Unit. At the Gatsby Unit, Hassabis met future DeepMind co-founder Shane Legg.
Founding and Acquisition of DeepMind
Demis Hassabis co-founded DeepMind in 2010 alongside Shane Legg and Mustafa Suleyman. DeepMind's stated mission was to "solve intelligence" and use intelligence "to solve everything else." Hassabis served as chief executive officer from its creation through August 2026.
Google acquired DeepMind in 2014, reportedly for between $400 million and $650 million, according to CNBC, as published by NBC Miami. DeepMind drew worldwide attention with AlphaGo. According to DeepMind's research records, AlphaGo defeated European champion Fan Hui 5-0 in October 2015 before beating Go champion Lee Sedol 4-1 in Seoul in March 2016, a match watched by more than 200 million people.
In April 2023, Google merged DeepMind with the Brain team from Google Research to form Google DeepMind, appointing Hassabis as CEO to lead general AI development as outlined on the Google Blog. On December 6, 2023, Pichai and Hassabis jointly announced Gemini 1.0 in Ultra, Pro, and Nano sizes on Google's blog, describing Gemini as the first realization of the vision behind forming Google DeepMind.
Protein Folding and the 2024 Nobel Prize
The 2024 Nobel Prize in Chemistry was awarded for work on protein design and protein structure prediction. As recorded by the Nobel Foundation, one half of the prize went to David Baker for computational protein design, and the other half was awarded jointly to Demis Hassabis and John Jumper for protein structure prediction.
The Nobel committee cited the AlphaFold2 system developed in 2020 by Hassabis and Jumper, stating in their press release that the model solved a 50-year-old challenge in biology. AlphaFold had demonstrated its predictive accuracy at the CASP14 competition in November 2020 by reaching a median score of 92.4 Global Distance Test (GDT) across all targets, as documented on the DeepMind blog. Hassabis received a one-fourth share of the prize with his official affiliation listed as Google DeepMind, London, on his Nobel facts page.
On May 8, 2024, Google DeepMind and Isomorphic Labs introduced AlphaFold 3, predicting interactions across proteins, DNA, RNA, and ligands, and released the model weights for academic use in November 2024 as detailed on the Google Technology Blog. Hassabis was appointed a Commander of the Most Excellent Order of the British Empire (CBE) in 2017 and was knighted in 2024 for his work in artificial intelligence.
Commercial Drug Discovery at Isomorphic Labs
Isomorphic Labs is an Alphabet company that applies AI to drug design. Alphabet announced the company on November 4, 2021, spinning it out of DeepMind under Hassabis's leadership, as covered by TechCrunch. Hassabis continues to serve as founder and chief executive officer of Isomorphic Labs, whose team page lists him as Sir Demis Hassabis.
In his August 2026 note to Google DeepMind staff, Hassabis wrote that he would lean into his role at Isomorphic Labs, reiterating his conviction that the primary application of AI should be to improve human health. Our Isomorphic Labs explainer covers how the company works.
In May 2026, Isomorphic Labs raised $2.1 billion in Series B funding led by Thrive Capital, according to an Isomorphic Labs press release. Hassabis called the round "a massive vote of confidence" in Isomorphic's "AI-first approach to drug design and development."
Definitions and Forecasts for Artificial General Intelligence
In a March 2025 interview reported by NBC News, Hassabis stated that AGI is "a system that's able to exhibit all the complicated capabilities that humans can," forecasting that those capabilities would start to emerge over the following five to ten years.
In his August 2026 note to Google DeepMind staff, Hassabis wrote that he has worked toward AGI his whole life and now feels "it is close at hand." In the accompanying announcement, Sundar Pichai wrote: "Demis has described us as standing in the foothills of the singularity."
For other leaders shaping the field, see our list of AI people to watch.
Executive Compensation and Net Worth Records
No major financial publication has published an official net worth valuation for Demis Hassabis as of October 2026. We found no net worth figure from Forbes, Bloomberg, or The Sunday Times.
Estimates that circulate online cite no named outlet. His compensation as Alphabet chief scientist has not been published.
Frequently Asked Questions
- Koray Kavukcuoglu is the Senior Vice President of Google DeepMind running day-to-day operations as of October 2026, reporting directly to Alphabet CEO Sundar Pichai. Demis Hassabis stepped back from daily operations on August 5, 2026, to serve as Chair of Google DeepMind and Chief Scientist of Alphabet.
- In March 2025, Demis Hassabis defined artificial general intelligence (AGI) as "a system that's able to exhibit all the complicated capabilities that humans can." In an August 2026 message to staff, he wrote that he feels AGI is "close at hand."
- Demis Hassabis received one fourth of the 2024 Nobel Prize in Chemistry alongside John Jumper "for protein structure prediction." The Nobel citation credits their development of AlphaFold2, an AI model that the Nobel committee said solved a 50-year-old problem: predicting proteins' complex structures.