The Scoop
Google DeepMind CEO Demis Hassabis’s departure Wednesday from the top job at the company he founded in 2010 and sold to Google in 2014 was at least a year in the making, said two people familiar with his thinking.
Hassabis had been drifting away from the day-to-day responsibilities of running the company’s Gemini AI models and its consumer AI strategy, said the people, increasingly shifting them to Koray Kavukcuoglu, the company’s chief AI architect.
Hassabis wasn’t pushed out against his will, said both people. Rather, he struggled to get satisfaction out of being in the role of a tech executive, rather than a visionary scientist.
Hassabis, who won a Nobel Prize in 2024, has recently been his most animated when talking about Isomorphic Labs, Google’s biotech spinout that he runs. His passions lie in using AI to solve scientific puzzles, like curing diseases and discovering new materials, and in ensuring that AI doesn’t accidentally cause catastrophic harm to humanity.
But the move, which coincided with the departure of chief scientist Jeff Dean, sent Google’s stock price down 4% Wednesday, and raised doubts about the company’s standing in the fast-paced AI race. Its AI models are roughly six months behind the frontier on coding ability, where most of the compute power is currently being consumed.
In its recent earnings on July 22, it reported revenue up 24% to $120 billion, and cloud revenue up 82%. But its capital expenditures, like just about every large tech company, are at record highs and pushed its free cash flow into negative territory for the first time since its 2004 IPO.
The feeling at Google, according to one executive, is that the management change will help accelerate the development of AI, rather than hold it back. While Hassabis had become the face of the company’s AI efforts, he wasn’t focused on the part of the company most associated with its standing in the race.
Dean has been a key contributor to Google’s underlying technology for more than two decades, and his departure was received with a sense of sadness by people at the company, including CEO Sundar Pichai, said a person familiar with the matter. But unlike Hassabis, Dean had few direct reports and wasn’t in charge of a major division.
In this article:
Know More
Kavukcuoglu is filling the role vacated by Hassabis at DeepMind. Now in charge of shepherding the much-anticipated Gemini 4 model, Kavukcuoglu joined Hassabis in DeepMind’s early days, having studied under AI pioneer Yann LeCun.
Hassabis’s departure from DeepMind also revives a question that’s been dogging Google all year: Can it still recruit top AI talent? The company recently lost John Jumper, who shared the Nobel with Hassabis, to Anthropic. The leadership change could beget more departures, as structural changes like this tend to do.
Reed’s view
I’ve interviewed Hassabis a handful of times since the ChatGPT moment. My most recent conversation with him was in June and it was clear something had shifted. Google had reoriented the company to improve the speed of deployment of AI products. And some of those products, like Notebook LM, have been huge successes. Gemini is the only chatbot that rivals OpenAI on the consumer front, and it’s gaining ground. But Hassabis seemed uninterested in talking about that stuff.
As one person said to me today, he’ll be a lot happier going forward and will probably win another Nobel in the process.
Google is still probably the best positioned company to win the consumer AI race, by virtue of its compute capabilities (thanks to its custom-built TPUs), its existing consumer base and its control of the world’s most popular mobile operating system.
Of course, that raises the question: Why hasn’t it already created some kind of AI super app that blows everyone out of the water. That criticism, though, assumes the world is changing faster than it is. Yes, AI capabilities have advanced to astonishing levels in recent years. But adoption is still barely occurring across the economy and globe.
That’s in part because there aren’t enough AI data centers and the costs are still too high. More compute will increase capabilities and make it less risky for Google to roll out services that today might malfunction in embarrassing and costly ways.
Google can’t afford to take its foot off the gas and it’s facing more competitive pressure than it has since its early days. The race is far from over and DeepMind still has plenty of tread on its tires.