Google DeepMind’s 3 Founders Are Now Shaping A.I. From Very Different Posts
The recent departure of several high-profile executives from Google DeepMind marks a distinct new era for the world-renowned research lab and a natural moment to trace the diverging paths of the trio who built it.
Long before it became an A.I. powerhouse, Google DeepMind was simply DeepMind. Founded in London by machine learning researchers Demis Hassabis, Shane Legg and Mustafa Suleyman, the lab was launched with the goal of achieving artificial general intelligence (AGI). Hassabis and Suleyman met as teenagers in London, while Hassabis later met Legg at University College London’s Gatsby Computational Neuroscience Unit.
Starting with a modest $1.85 million in funding from Peter Thiel’s Founders Fund, which took a 25 percent equity stake, DeepMind proved its early mettle by pioneering deep reinforcement learning, famously teaching an A.I. to master seven Atari games from scratch with zero prior instructions.
In 2014, Google acquired DeepMind for £400 million ($650 million). The tech giant merged its in-house A.I. research unit, Google Brain, with DeepMind to form Google DeepMind, helmed by Hassabis.
Last month, Google DeepMind announced that Hassabis would step back from the day-to-day operations and shift into the role of chairman of Google DeepMind while taking on the title of Alphabet chief scientist, replacing Jeff Dean. The standalone CEO title was abolished. In the meantime, Longtime DeepMind chief technology officer Koray Kavukcuoglu was promoted to the lab’s senior vice president, reporting directly to Alphabet CEO Sundar Pichai.
Today, the original three founders of DeepMind find themselves steering the A.I. revolution from three different vantage points:
Demis Hassabis, 50
Current Role: CEO and Chair of Google DeepMind; CEO of Isomorphic Labs
Hassabis remains a central figure in Google’s A.I. playbook as chairman of Google DeepMind. Meanwhile, he also runs Isomorphic Labs, the Alphabet-backed A.I. drug discovery spinout he launched in 2021. His work is split between frontier AGI research and biotech breakthroughs. In 2024, Hassabis, along with Google DeepMind vice president John Jumper and biochemist David Baker, received the Nobel Prize in chemistry for developing an A.I. model called AlphaFold2 that’s able to predict the structure of virtually all proteins.
In his new role as Alphabet’s chief scientist, “I will continue to work closely with Sundar on strategic and global AGI matters, and to advise [DeepMind] leads, from our awesome new London Platform 37 offices,” he said in a statement last month. He added that he will be leaning further into Isomorphic Labs, as he finds improving human health to be the top priority of A.I.
Mustafa Suleyman, 42
Current Role: CEO of Microsoft AI
Suleyman was Google DeepMind’s head of applied A.I. He notably launched DeepMind Health, now part of Google Health, and applied the lab’s algorithms to cut Google’s data center cooling costs by 40 percent. During his tenure, he also established the DeepMind Ethics & Society unit to analyze the real-world impacts of A.I. and help researchers implement ethical standards.
He left Google in 2022 to co-found personal A.I. assistant Inflection AI with Reid Hoffman. In March 2024, he was tapped by Satya Nadella to head Microsoft AI, leading Microsoft’s consumer A.I. products like Copilot.
Today, his work focuses on “humanist superintelligence” and A.I. safety frameworks. He recently authored Microsoft’s draft A.I. Code of Conduct to ensure frontier models remain subordinate to human control.
Shane Legg, 53
Current Role: Chief AGI Scientist at Google DeepMind
The quiet third founder of DeepMind, Legg has stayed true to the lab’s original charter. He leads Google DeepMind’s technical AGI safety teams, serves on the nominating committee for the AI Safety Foundation, and focuses heavily on talent acquisition and long-term AGI roadmaps. In January, Legg said he sees a 50-percent chance of reaching “minimal AGI,” where A.I. handles standard human cognitive tasks, by 2028.
“I think there are many problems in the world that could be helped by having an extremely capable and ethical intelligence system,” Legg said in an interview for his TIME100 AI placement in 2023. “The world could become a much, much better place.”
