Google's AI Crisis: DeepMind Loses Independence and Hassabis Is as Good as Gone

According to a Guardian report, a former Google manager sees 'the era of DeepMind as an independent player' as over, and a current DeepMind employee calls the lab 'just another subdivision of Google.' AI researcher Koray Kavukcuoglu takes over operational leadership from the US, without assuming Hassabis's CEO title. Per Pathfounders, Hassabis wanted to leave simultaneously with Jeff Dean, but Google feared a stock drop, so he was promoted to Chairman instead. He plans to focus more on his scientific work and the drug startup Isomorphic Labs, and could leave DeepMind in the coming months. All AI development around Gemini is moving to the Bay Area, and Sergey Brin is expected to take on more responsibility. Brin had already come out of retirement during Google's first 'red alert' over ChatGPT and got involved in development, leading to the merger of Google Brain and DeepMind into 'Google DeepMind.' Two perspectives exist on the situation: SemiAnalysis argues Google is losing the race for frontier AI models due to years of overly cautious compute allocation and a bureaucratic, risk-averse culture, comparing it to IBM and Intel's mistakes, while Tim O'Reilly argues Google is intentionally competing on infrastructure rather than technological edge, citing Thomas Edison vs. George Westinghouse. However, recent events suggest Google wanted but could not catch up at the frontier, especially in coding, and Gemini 3.1 Pro remains in preview, while 3.5 Pro has been delayed.