admin The shift places Koray Kavukcuoglu, Alphabet’s chief AI architect, at the centre of the company’s effort to turn advanced research into widely used products. Kavukcuoglu has moved full-time from London to California after previously dividing his schedule between the two locations, bringing him closer to Alphabet chief executive Sundar Pichai and senior leaders across Google’s businesses.
Google is also gathering parts of its Gemini model operation in Mountain View. The post-training team, which refines models after their initial large-scale training, is becoming more concentrated at the company’s headquarters rather than being distributed across several offices. The move is intended to reduce organisational friction and shorten the time needed to move research breakthroughs into Search, Cloud, Workspace, Android and other services.
Kavukcuoglu was appointed chief AI architect in 2025 with a mandate to improve the integration of Google’s models across its products. His authority has since expanded as the company unified more of its model development under Google DeepMind and sought to establish a common technical foundation for its generative AI services.
He now oversees thousands of researchers and engineers working on Gemini and has assumed day-to-day responsibility for Google DeepMind following a broader leadership overhaul. Demis Hassabis has moved from the unit’s chief executive role to become Alphabet’s chief scientist and chairman of Google DeepMind, allowing him to focus on long-term research, artificial general intelligence and the wider social consequences of advanced systems.
The changes reflect the pressure facing Google as generative AI reshapes its core markets. OpenAI established an early lead in consumer chatbots through ChatGPT, while Anthropic’s Claude has gained ground among businesses and software developers. Both rivals are headquartered in San Francisco, placing their senior teams within reach of Silicon Valley’s research talent, investors and enterprise customers.
OpenAI has also expanded into Google’s backyard by leasing a five-building office complex in Mountain View with space for more than 1,000 employees. The move adds another layer to the competition for engineers and researchers in an area already dominated by Alphabet, Meta, Apple and a growing cluster of AI start-ups.
Google retains significant advantages, including its global search platform, YouTube, Android, Chrome, cloud infrastructure and custom-designed tensor processing units. Its challenge has been coordinating those resources quickly enough to compete with companies built around a narrower AI mission.
Kavukcuoglu’s leadership style places greater emphasis on benchmarks, product delivery and collaboration with commercial divisions. He has worked closely with Google Cloud chief executive Thomas Kurian and executives responsible for Search and other major products. Teams involved in training Gemini within the search organisation have also been moved closer to the central AI structure.
The strategy is designed to limit duplication across Google. Product teams have historically adapted AI models for their own requirements, creating different versions and development processes. Google DeepMind is working towards a model architecture that can be used more consistently across the company, reducing the need for separate systems and making upgrades easier to deploy.
The centralisation comes as Google manages the loss of several influential researchers. Jeff Dean, Sanjay Ghemawat, Oriol Vinyals and Quoc Le have left to establish Discovery Loop, a public benefit corporation focused on machine learning, science and engineering. Google is investing in the venture and providing computing capacity through its cloud business.
Other departures have strengthened competitors. Noam Shazeer, a leading figure in the development of Gemini, joined OpenAI, while Nobel Prize-winning researcher John Jumper moved to Anthropic. The exits have intensified scrutiny of Google’s ability to retain specialised talent as compensation packages and research budgets rise across the industry.
Alphabet shares fell after the leadership changes were announced, reflecting investor concerns about execution, spending and delays affecting the next flagship Gemini model. The company has committed tens of billions of dollars to data centres, chips and AI infrastructure while facing pressure to demonstrate that the investment can generate sustainable revenue.
