Demis Hassabis Steps Down as DeepMind CEO, Moves to AGI Focus

Demis Hassabis stepped down as CEO of Google DeepMind on August 5, 2026, taking on a chairman and Alphabet chief scientist role to focus on AGI strategy and Isomorphic Labs. The announcement came the same day Jeff Dean, Sanjay Ghemawat, Oriol Vinyals, and Quoc Le revealed they were leaving Google to co-found Discovery Loop. Alphabet shed roughly 5 percent of its market value on the news.

6 min read
Demis Hassabis, DeepMind CEO leadership transition and AGI focus, 2026
Demis Hassabis at the 2024 Nobel Prize week in Stockholm, where he collected the Chemistry prize for AlphaFold.· Photo by Arthur Petron, via Wikimedia Commons (CC BY-SA 4.0)

Demis Hassabis on August 5 gave up day-to-day leadership of Google DeepMind, handing the unit's top operational role to chief technology officer Koray Kavukcuoglu and moving himself to chairman of the unit and chief scientist of parent company Alphabet, Bloomberg reported. The same day, Jeff Dean, Sanjay Ghemawat, Oriol Vinyals, and Quoc Le disclosed they were leaving Google to co-found Discovery Loop, a public-benefit corporation focused on AI-accelerated scientific research, according to TechCrunch. Alphabet shares fell roughly 5 percent, erasing an estimated $160 billion to $200 billion in market value.

What the restructuring means for Gemini

Kavukcuoglu becomes senior vice president of Google DeepMind, reporting directly to Sundar Pichai rather than through Hassabis, giving the Gemini product and research operation a more direct line to Alphabet's executive leadership, Axios reported on August 5. He had already been absorbing day-to-day responsibilities from Hassabis over the past year, including directing model development and presenting major Gemini releases publicly.

Hassabis told staff his motivation was proximity to AGI. "I've been working towards AGI my whole life and now, like many of you, I feel it is close at hand," he wrote in an internal memo cited by CNBC. In his new roles, he will focus on AI safety, policy, and strategic questions about what AGI's arrival means for society, while also deepening his involvement with Isomorphic Labs, Google's AI drug discovery subsidiary.

StartupHub.ai data shows Google DeepMind operating with 8,831 employees and $1.33 billion in annual revenue. The decision to move Hassabis out of operational leadership of an organization that size, while Gemini competes at the frontier against OpenAI and Anthropic, is a structural bet that separating research vision from product execution will sharpen both.

The talent departure layered beneath the headline

The restructuring was announced on the same day as one of the largest single-event talent departures from Google's AI division in the company's history. Jeff Dean, a 27-year Google veteran and principal architect of TensorFlow, Google Brain, and the company's large-scale ML infrastructure, is leaving alongside Sanjay Ghemawat, Oriol Vinyals, and Quoc Le to co-found Discovery Loop, TechCrunch confirmed. Discovery Loop's stated mission is using AI to accelerate scientific research across ML automation, hardware design, drug discovery, and clean energy.

The August exits compound a pattern that started in June, when John Jumper, who shared the 2024 Nobel Prize with Hassabis for AlphaFold, left Google DeepMind for Anthropic. Together, these five researchers represent the highest concentration of senior AI talent to leave a single lab in 2026, CNBC noted.

The departures sit against a specific competitive context: Gemini 3.5 Pro has been months behind schedule even as OpenAI and Anthropic shipped major model updates in 2026, per Axios. Bloomberg's August 6 analysis noted Kavukcuoglu inherits both the model delay and a research team now smaller by four of its most recognized names.

What Hassabis is building at Isomorphic Labs

Hassabis's continued leadership of Isomorphic Labs is the clearest signal of where his attention is moving. The company, which uses AlphaFold-derived models to identify novel drug candidates, has raised $4.7 billion in total funding with its most recent round closing in May 2026, according to StartupHub.ai data. Among the AI drug discovery companies tracked by StartupHub.ai, the next-largest disclosed funding stands at $28 million; Isomorphic Labs is capitalised at a ratio of roughly 168 to 1 relative to the nearest tracked peer, which reflects the scale of Alphabet's commitment to the category.

Isomorphic Labs runs active drug discovery partnerships with Eli Lilly and Novartis, using AI models to identify candidate molecules at a pace Hassabis has described as orders of magnitude faster than conventional wet-lab methods. Moving to the chairman and chief scientist roles at DeepMind frees him to spend more time on that operation directly, rather than managing Gemini model timelines and commercial product priorities.

The question his restructuring leaves open is whether one executive can hold together chairman of an 8,831-person AI lab, chief scientist of a $2 trillion public company, and CEO of an AI drug discovery startup with $4.7 billion in backing. The separation of Gemini's operational management from his oversight removes one source of that tension; the broader concentration of responsibility in a single person remains.

What it means

The August 5 announcements collectively describe a Google reorganizing its AI leadership under competitive pressure while losing some of the researchers who built its foundational systems. Hassabis's move to chairman gives him the cleaner mandate he said he wanted: AGI safety and strategy rather than quarterly model shipping. Whether Kavukcuoglu's direct line to Pichai is enough to close the gap on Gemini is the immediate operational question. The longer strategic issue is whether an AI drug discovery company with $4.7 billion in backing and a Nobel-winning chief executive, running in parallel to a large public lab, is a feature of Google's AI strategy or a competing claim on its most valuable resource.

Sources

Editorial standards: every claim is sourced. Tips: [email protected]

© 2026 StartupHub.ai. All rights reserved. Do not enter, scrape, copy, reproduce, or republish this article in whole or in part. Use as input to AI training, fine-tuning, retrieval-augmented generation, or any machine-learning system is prohibited without written license. Substantially-similar derivative works will be pursued to the fullest extent of applicable copyright, database, and computer-misuse laws. See our terms.