Updated November 2026 · his new Google role, Isomorphic Labs and AlphaFold 3 added; facts rechecked
The short version: Demis Hassabis (born 1976, London) is a former chess prodigy and video game designer who co-founded DeepMind in 2010. His teams built AlphaGo, which beat a world Go champion, and AlphaFold, which predicts the shapes of proteins and won him a share of the 2024 Nobel Prize in Chemistry. Since August 2026 he has been chair of Google DeepMind and chief scientist of Alphabet, Google’s parent company, and he runs the drug-discovery company Isomorphic Labs.
How a chess prodigy and game designer ended up building some of the most important AI in science.
Demis Hassabis has had several careers. He was a chess master as a boy, a hit game designer as a teenager, a neuroscientist by 30 and the co-founder of one of the world’s leading artificial intelligence labs.
Demis Hassabis in eight steps
- 1994Co-designs the hit game Theme Park at Bullfrog, aged 17, before starting at Cambridge
- 1998Founds Elixir Studios after a computer science degree and a spell at Lionhead
- 2009Finishes a PhD in cognitive neuroscience at University College London
- 2010Co-founds DeepMind in London with Shane Legg and Mustafa Suleyman
- 2016AlphaGo beats Lee Sedol, one of the world’s best Go players, 4 games to 1
- 2020AlphaFold 2 predicts protein shapes about as well as lab experiments
- 2024Shares the Nobel Prize in Chemistry with John Jumper and David Baker
- 2026Becomes chair of Google DeepMind and chief scientist of Alphabet
Words you will meet
- Protein
- A molecule made from a chain of amino acids that folds into a 3D shape; the shape decides its job in the body.
- Reinforcement learning
- A way of training AI by trial and error, rewarding it when it does well, as with AlphaGo.
- AGI
- Artificial general intelligence: AI that could do most thinking tasks a person can. It does not exist yet.
- Chief scientist
- A senior research role. At Alphabet, Google’s parent company, Hassabis now advises on long-term science rather than running a lab.
Who is Demis Hassabis?
Hassabis co-founded DeepMind in 2010 and led it for 16 years, through Google’s purchase in 2014 and the 2023 merger that created Google DeepMind. On August 5, 2026, Google’s chief executive Sundar Pichai announced that Hassabis would become chair of Google DeepMind and chief scientist of Alphabet, while continuing to lead Isomorphic Labs. Koray Kavukcuoglu, a long-time DeepMind leader, now runs Google DeepMind day to day, including work on the Gemini models. Hassabis also holds a knighthood, awarded in March 2024 “for services to Artificial Intelligence.”
A chess prodigy in north London
Hassabis was born on July 27, 1976, in London, to a Greek-Cypriot father and a Chinese-Singaporean mother, according to the Nobel Foundation. He grew up in Finchley, in north London, the eldest of three children; his father was a toy salesman and his mother worked at the department store John Lewis (The Times Magazine).
He learned chess at four and reached a master-level rating of 2300 at 13. At one point he was the second-highest-rated under-14 player in the world, behind Judit Polgar (ChessBase). He has said chess taught him to plan many moves ahead and to think about how thinking works, the question that would occupy the rest of his career.
From Theme Park to Cambridge, and back to games
At 17 Hassabis joined Bullfrog Productions, the British studio run by Peter Molyneux, and co-designed Theme Park, a best-selling simulation game (MCV). He then took up his place at Queens’ College, Cambridge, and graduated in computer science in 1997 with a double first, the top grade (his CV).
After a short spell at Molyneux’s next studio, Lionhead, he founded Elixir Studios in 1998. It made Republic: The Revolution and Evil Genius, ambitious simulation games with AI-driven characters, and closed in 2005. Building simulated worlds showed him, he later said, how simple rules can produce complex behavior.
Why did he study the brain?
Hassabis believed the fastest route to general AI ran through understanding the human brain. In 2005 he started a PhD in cognitive neuroscience at University College London with Eleanor Maguire, a leading memory researcher, and finished in 2009. His best-known paper, published in 2007, showed that people with damage to the hippocampus, a brain area key to memory, also struggle to imagine new experiences, suggesting memory and imagination share the same machinery.
How did DeepMind start?
In 2010 Hassabis co-founded DeepMind Technologies in London with Mustafa Suleyman, a childhood friend, and Shane Legg, a researcher he met at UCL. The stated mission: solve intelligence, then use it to solve everything else. Read more about Suleyman in our Mustafa Suleyman profile.
DeepMind’s first big result came in a 2013 workshop paper: one learning program played seven Atari games using only the screen pixels and the score, and beat a human expert on three of them. A 2015 follow-up in Nature extended this to 49 games. The method, called deep reinforcement learning, learns by trial, error and reward.
Google bought DeepMind in 2014 for about £400 million, its biggest acquisition in Europe at the time. Reports then said the founders asked for an AI ethics board as part of the deal, though its membership was never made public.
AlphaGo and the Lee Sedol match
Go, an ancient board game with more possible positions than there are atoms in the universe, was long thought to be decades beyond computers. In March 2016 DeepMind’s AlphaGo beat Lee Sedol, winner of 18 world titles, by four games to one in Seoul. More than 200 million people watched worldwide. In the second game, AlphaGo’s move 37 looked like a mistake to experts, and it went on to win that game.
The next version, AlphaGo Zero, learned only by playing itself, with no human games to study. After three days it beat the version that had defeated Lee Sedol 100 games to 0, and after 40 days it was stronger than every earlier version. AlphaZero then applied the same approach to chess and shogi, a Japanese form of chess.
What is AlphaFold, and why did it win a Nobel Prize?
Proteins are the tiny machines that do most of the work in living things, and a protein’s 3D shape decides what it does. Working out that shape in a lab can take years. Predicting it from the protein’s recipe of amino acids had been an open problem in biology for about 50 years.
DeepMind’s AlphaFold 2, revealed in 2020, predicted protein shapes about as accurately as lab methods for many proteins. The team, led by John Jumper, then released predictions for over 200 million proteins, nearly every one known to science, in a free AlphaFold database run with the European Bioinformatics Institute (EMBL-EBI). According to the Nobel Committee, AlphaFold 2 has been used by more than two million people in 190 countries.
In October 2024 the Nobel Prize in Chemistry went half to David Baker “for computational protein design” and half jointly to Hassabis and Jumper “for protein structure prediction.” Earlier that year, Google DeepMind and Isomorphic Labs released AlphaFold 3, which predicts how proteins interact with DNA, RNA and drug-like molecules, with a free AlphaFold Server for researchers. For what this means for medicine, see our guide to AI in drug discovery.
Isomorphic Labs: turning AlphaFold into medicines
Hassabis founded Isomorphic Labs in 2021, a sister company to DeepMind that uses AI to design new drugs, and he is its chief executive. In January 2024 it signed deals with Eli Lilly and Novartis worth about $3 billion in potential payments, and in May 2026 it raised $2.1 billion in new funding. Its first medicines have not yet reached clinical trials; the company has said it hopes to start them by the end of 2026.
What does Hassabis do at Google now?
Google merged DeepMind with its Google Brain team in 2023 to form Google DeepMind, with Hassabis as chief executive. The combined lab builds Gemini, Google’s AI model family; the current flagship is Gemini 4 Argon, announced in September 2026. It also keeps up a steady run of science projects, including AlphaEvolve, which discovers new algorithms, the Genie 3 world model, which generates explorable 3D worlds, and AlphaGenome, which predicts how changes in DNA affect genes. Our Google DeepMind guide covers the lab in more depth.
In his new role he focuses on long-term research. In Pichai’s announcement, Hassabis described the field as “standing in the foothills of the singularity.”
What does Hassabis think about AI safety and AGI?
Hassabis signed the 2023 Statement on AI Risk, which says reducing the risk of extinction from AI should be a global priority alongside pandemics and nuclear war. That October he told The Guardian: “We must take the risks of AI as seriously as other major global challenges, like climate change.” He has called for an international scientific panel on AI, a research collaboration like CERN for AI safety, and eventually an equivalent of the International Atomic Energy Agency.
In July 2026 he went further, proposing in a written plan that the US set up a watchdog to test the most powerful AI models before release, modeled on FINRA, the body that oversees Wall Street brokers. Our daily newsletter explained the plan in its July 14, 2026 edition.
His view of when artificial general intelligence (AI that can do most things a person can) might arrive has moved closer. In 2025 he told 60 Minutes it could come within five to ten years, and in 2026 he has suggested it may be sooner.
Unlike most AI leaders, Hassabis is best known for a scientific tool that biologists use every day, not a chatbot. If you want one thing to try, search for a protein you have heard of, like insulin, in the free AlphaFold database and spin its predicted shape in your browser.
Frequently asked questions
What did Demis Hassabis win the Nobel Prize for?
For AlphaFold, the AI system that predicts the 3D shapes of proteins. He shared half of the 2024 Nobel Prize in Chemistry with John Jumper “for protein structure prediction”; the other half went to David Baker for designing new proteins with computers.
Is Demis Hassabis still the CEO of Google DeepMind?
No. Since August 5, 2026, he has been chair of Google DeepMind and chief scientist of Alphabet. Koray Kavukcuoglu runs the lab day to day. Hassabis is still chief executive of Isomorphic Labs.
What is DeepMind’s relationship with Google?
Google bought DeepMind in 2014. In 2023 it merged DeepMind with Google Brain to form Google DeepMind, which builds the Gemini models and runs much of Google’s AI research.
What is AlphaFold?
AlphaFold is an AI system from Google DeepMind that predicts a protein’s 3D shape from its sequence of amino acids. Its predictions for over 200 million proteins are free in the AlphaFold database, and AlphaFold 3 also models how proteins interact with DNA, RNA and drugs.
Does Demis Hassabis believe AGI is coming?
Yes. He founded DeepMind to build general intelligence, and his estimates have shortened over time, from five to ten years in 2025 to possibly sooner in 2026. He also argues it needs international safety rules and testing before powerful models are released.
For the wider story, see our guide to the AI pioneers. New to AI yourself? Start with what artificial intelligence is and how to use it, then try a few of the best AI tools for beginners, browse our AI tools directory or take one of these free AI courses.
Made for regular people, by a real person
A free daily newsletter on the people and labs shaping AI, and what their work means for you. Curated and edited by a human, in plain English.
Free forever. Unsubscribe anytime.
Sources
- Google: The next chapter of our AI momentum (Sundar Pichai, 5 August 2026)
- Nobel Prize in Chemistry 2024 press release
- Nobel Foundation: Demis Hassabis facts
- Academia Europaea: Demis Hassabis CV
- The Times Magazine profile of Demis Hassabis (PDF)
- ChessBase: BBC’s Across the Board with Demis Hassabis
- MCV: Theme Park lead programmer sells AI start-up to Google
- Hassabis et al.: Patients with hippocampal amnesia cannot imagine new experiences, PNAS (2007)
- Mnih et al.: Playing Atari with Deep Reinforcement Learning (2013)
- Google DeepMind: AlphaGo
- Google DeepMind: AlphaGo Zero, starting from scratch
- AlphaFold Protein Structure Database (EMBL-EBI)
- Google: AlphaFold 3 (May 2024)
- TechCrunch: Isomorphic signs deals with Eli Lilly and Novartis (January 2024)
- Isomorphic Labs: funding announcement (May 2026)
- Google DeepMind: announcing Google DeepMind (2023)
- Google AI for Developers: Gemini 4 Argon
- Google DeepMind: AlphaEvolve
- Google DeepMind: Genie 3
- Google DeepMind: AlphaGenome
- The London Gazette: knighthood notice (March 2024)
- Center for AI Safety: Statement on AI Risk (2023)
- The Guardian: Hassabis on AI risk and climate change (October 2023)
- The Verge: Hassabis proposes an AI watchdog (July 2026)
- CBS 60 Minutes: Demis Hassabis interview (2025)
- Demis Hassabis on Grokipedia

