Demis Hassabis Steps Back From Google DeepMind CEO Role as Google Reshapes Its AI Leadership

Google is making one of its biggest changes to its artificial intelligence leadership since creating Google DeepMind.

Demis Hassabis, the co-founder and long-time leader of DeepMind, is stepping away from the company's day-to-day CEO responsibilities. He will become Chair of Google DeepMind and Chief Scientist of Alphabet, allowing him to focus more heavily on long-term AI research, artificial general intelligence (AGI) and the wider societal implications of increasingly powerful AI systems.

Hassabis is not leaving Google. The change moves him from operational leadership toward a broader research and strategic role. Koray Kavukcuoglu, previously Google DeepMind's chief technology officer, will take responsibility for running the organisation on a day-to-day basis.

Why is Hassabis changing roles?

He has argued that AI is approaching a pivotal stage and that AGI could be much closer than previously expected. In his new role, Hassabis will have more freedom to concentrate on the longer-term scientific and societal questions surrounding advanced AI rather than managing daily operations of one of the world's largest AI laboratories.

Alphabet CEO Sundar Pichai said the two had been discussing a position that would allow Hassabis to focus his attention on shaping the future of AGI. This is therefore better understood as a change in focus rather than an exit.

Google DeepMind is getting a new operational leader

Reuters reports that Kavukcuoglu has already been more directly involved than Hassabis in collaborations between DeepMind and Google Cloud, and his appointment is expected to push Gemini toward a more product-oriented direction.

Google DeepMind isn't simply a research laboratory anymore. Its technology sits at the centre of Google's consumer products, cloud business and enterprise AI strategy, so the company must both advance the science of AI and turn that research into competitive products.

The timing is significant

The change comes amid intense competition from OpenAI and Anthropic, while Chinese AI companies also put pressure on leading American laboratories. Reuters reports that Google's next major Gemini upgrade had not yet launched despite an earlier planned release, contributing to concerns among some investors. Alphabet's shares fell about 4% following the announcement.

Google has nevertheless made significant progress. Its Gemini family has become one of the most important competitors to OpenAI's models, and Google retains major advantages in computing infrastructure, search, cloud services, data and access to billions of users.

Hassabis is also focused on AI safety and AGI

AI Magazine reports that Hassabis recently warned about potential risks involving cybersecurity, nuclear weapons and biological threats as AI capabilities advance, and argued for stronger mechanisms for testing and evaluating increasingly capable AI systems.

In June 2026, Google DeepMind published an AI Control Roadmap focused on securing internal systems against more capable AI agents, and a research publication examining possible pathways from AGI toward artificial superintelligence, including scaling, new paradigms and recursive improvement.

Google's AI leadership is changing beyond Hassabis

Several prominent Google AI and engineering leaders are also leaving. Jeff Dean, Sanjay Ghemawat, Oriol Vinyals and Quoc Le are departing to establish Discovery Loop, a new public-benefit corporation focused on research in machine learning, science and engineering. Google therefore isn't simply changing one executive — it is undergoing a broader restructuring of its AI leadership and research organisation.

What is Discovery Loop?

Discovery Loop is being created around the idea of using AI to accelerate scientific and engineering research. Reuters reports that the new company has received investment from Google and will also have a partnership with Google Cloud for computing capacity — an unusual situation in which Google loses senior researchers while maintaining a relationship with their new company.

Is Google falling behind?

It would be too simplistic to say the role change means Google is losing the AI race. Google DeepMind remains one of the world's most powerful AI research organisations, with breakthroughs including AlphaGo and AlphaFold. But the competitive environment has changed dramatically, and even the largest AI laboratories cannot afford to slow down.

A shift from AI research to AI execution?

The division of responsibilities is one of the most interesting aspects of the change. Hassabis can focus on AGI research, long-term AI strategy, AI safety, scientific discovery and societal impact. Kavukcuoglu can focus on Gemini, product development, engineering, Google Cloud integration, commercialisation and operational execution.

That could allow Google to pursue two tracks simultaneously: working toward increasingly capable AI and AGI over the long term, while making Gemini and Google's AI products more competitive in the market today.

What does this mean for businesses?

Google's AI technologies increasingly appear across Workspace, Google Cloud, Search, Android, advertising, customer service, software development, data analytics and business automation. If Google's AI development accelerates, businesses could gain access to more capable models and agents through products they already use.

For businesses, the important question isn't whether Google wins against OpenAI or Anthropic. It is whether AI capabilities continue becoming cheaper, more capable and easier to integrate into everyday workflows.

What could it mean for Kenya?

Kenyan businesses are increasingly using cloud software, digital payments, online marketing, customer-service platforms and AI tools. More capable AI systems could enable SMEs to automate:

• Customer enquiries
• Marketing
• Data analysis
• Website content
• Software development
• Document processing
• Customer support
• Internal reporting
• Sales follow-ups

Google's large cloud and AI ecosystem could also make it easier for African businesses and developers to build AI-powered applications without developing foundation models themselves — particularly important for smaller companies that cannot afford to build sophisticated AI infrastructure from scratch.

Our take

Hassabis stepping away from the CEO position should not be interpreted as him stepping away from AI. In many ways, the opposite may be true: he is moving into a position where he can focus more heavily on the technology's long-term direction while another senior leader executes Google's commercial and engineering strategy.

For businesses, developers and technology leaders, Google's next moves will be worth watching closely. The company isn't simply competing to build the next chatbot — it is competing to define a significant part of the infrastructure, products and AI systems that could shape the next generation of the internet.

Sources: Reuters reporting on Google's AI leadership restructuring; Google DeepMind's official information on its mission, research and AI safety work; and AI Magazine's original report on Hassabis' role change.