Google sends new DeepMind chief to catch up with OpenAI and Anthropic

Koray Kavukcuoglu, the new head of Google DeepMind, faces the challenge of closing the gap with OpenAI and Anthropic in the AI race. Recently, Anthropic and OpenAI released new systems that received high praise, while Google has not released advanced models since early 2026. Kavukcuoglu, previously DeepMind's CTO and Google's chief AI architect, will take the role of senior vice president, succeeding co-founder and CEO Demis Hassabis, who becomes chairman of the board. Since OpenAI launched ChatGPT in 2022, Google has been playing catch-up, briefly taking the lead in October with Gemini 3. In 2026, Anthropic released the revolutionary model Mythos, and OpenAI's GPT-5.6 also received high marks. Google unveiled Gemini 3.1 Pro in February but has not released advanced models since. The company has formed a Code Strike department for AI coding models, but still trails OpenAI and Anthropic because it focused on monetizing technology, according to CNBC, though Google disputed that explanation. Google noted that progress toward AGI requires advances in AI agents, coding, robotics, and world models, citing DeepMind's achievements in robotics, video, and computer control. Kavukcuoglu joined DeepMind in 2012, before its acquisition by Google. Reporting directly to CEO Sundar Pichai, he will oversee Gemini models, advanced AI research, and AI developer teams. His promotion signals Google's shift towards implementation over deep research, experts say, as it will release new models and experimental products faster. Alphabet shares recently fell when Bloomberg reported a delay in Gemini 3.5 Pro. Despite falling behind OpenAI and Anthropic, experts believe it is premature to talk about a crisis at Google, as Kavukcuoglu may solve the challenge given that Google excels at product release and execution.