Google is centralising control of its artificial intelligence operations in Silicon Valley in a sweeping leadership overhaul that strengthens the influence of co-founder Sergey Brin and marks a significant shift in the company’s AI strategy. According to a dozen people familiar with the company, including current and former employees, the reorganisation consolidates decision-making in California as Google accelerates efforts to commercialise its Gemini family of AI models and compete more aggressively with rivals such as OpenAI and Anthropic.
As part of the restructuring, Sir Demis Hassabis is stepping back from the day-to-day management of Google DeepMind, the AI laboratory he founded and sold to Google for £400 million in 2014. Operational responsibility for the business will pass to Koray Kavukcuoglu, the former DeepMind chief technology officer, who relocated from London to Google’s headquarters in Mountain View last year and now leads development of the Gemini models from alongside Brin.
The leadership changes represent a major transformation at the US$4.4 trillion technology company, where commercial priorities are increasingly taking precedence over the research-driven culture that characterised DeepMind for more than a decade. Google is under mounting pressure to demonstrate that it can translate its extensive AI research into commercially successful products while addressing investor concerns over the cost of its AI investments and competing more effectively in enterprise AI and coding technologies.
People familiar with the reorganisation said the transition had been planned for several months, with Hassabis gradually handing operational responsibility for Gemini development to Kavukcuoglu as Google intensified work on its flagship AI systems. At the same time, Brin, who had largely withdrawn from daily operations for several years, has resumed an influential role in shaping the company’s AI strategy following the emergence of OpenAI’s ChatGPT. Despite holding no formal executive position, he has become increasingly involved in key decisions surrounding Gemini’s development.
Hassabis will assume the roles of Alphabet’s chief scientist and chair of Google DeepMind, allowing him to concentrate on long-term AI research while relinquishing overall responsibility for managing the business. Although he is widely respected for his scientific leadership, several people familiar with Google’s internal discussions said senior executives had become frustrated by what they viewed as his comparatively limited focus on the commercial demands of the company’s AI operations.
One example cited by a person familiar with the matter was Hassabis’s decision to make AlphaFold, the protein structure prediction system that earned him the 2024 Nobel Prize in Chemistry, freely available. While regarded as a landmark scientific achievement, the project generated little commercial return despite substantial investment. However, another person close to Google disputed suggestions of internal tensions, noting that Hassabis led the initial launch of Gemini and remained fully committed to its success.
Yann LeCun, a leading AI researcher and former chief AI scientist at Meta, said Hassabis’s move away from management reflected a common path taken by senior scientists. He said researchers with unique expertise could often contribute more effectively by pursuing research directly and inspiring others rather than overseeing large organisations.
The restructuring has unsettled employees within DeepMind’s London headquarters. Several current and former staff members said the changes had sent shockwaves through the organisation, with some fearing they marked the end of the research-focused culture that Hassabis had long championed. One former DeepMind executive now working for a rival company said multiple employees had already made contact expressing interest in leaving.
The overhaul comes as Google faces wider challenges in retaining senior AI talent. Jeff Dean, a highly regarded AI scientist and a 27-year veteran of the company, also announced his departure this week. He is launching a new AI start-up, Discovery Loop, with three colleagues while retaining Google’s backing. Investors also reacted cautiously to the leadership changes, with Alphabet shares falling five per cent following the announcements amid concerns over the escalating costs of Google’s AI ambitions and the company’s ability to convert research excellence into profitable products.
DeepMind has been central to Google’s AI strategy since chief executive Sundar Pichai merged it with Google Brain in 2023, creating a single organisation responsible for Gemini, AI safety, research and related applications. Hassabis described the expanded division as Google’s “AI engine room”, responsible for developing its family of Gemini large language models and other frontier AI systems.
Over the past year, Hassabis has also devoted increasing attention to Isomorphic Labs, the AI drug discovery company he founded, while emerging as one of the technology industry’s leading public voices on artificial intelligence. A friend said the new arrangement allowed him to retain significant influence without relocating to the United States and provided greater freedom to focus on Isomorphic Labs.
Although Google enjoyed renewed momentum following the release of its Gemini 3 models and the popularity of its Nano Banana AI image-editing tool, company executives remain concerned about its relative position in enterprise AI and coding models, where OpenAI and Anthropic have established an early advantage. Explaining the reorganisation to staff, Hassabis said he wanted to focus on the “big picture” and help shape the development of increasingly powerful artificial general intelligence, which he described as “close at hand”. Sundar Pichai said the leadership changes were designed to accelerate product development, telling employees that Google must remain focused on “the AI frontier” and continue building on its recent momentum.
