China’s AI Models Narrow the Gap With the West, Says Google DeepMind CEO

China’s artificial intelligence capabilities are rapidly catching up to those of the United States and other Western nations, according to Demis Hassabis, CEO of Google DeepMind. Speaking on a CNBC podcast, Hassabis said that leading Chinese AI models are now only “a matter of months” behind their Western counterparts, highlighting a significant acceleration compared to the gap seen one or two years ago.

Hassabis, who oversees Google’s AI strategy including its Gemini AI models and assistant, noted that while Chinese developers have made impressive progress, they have yet to deliver true breakthroughs that push beyond today’s frontier AI models. He questioned whether current approaches in China, as well as globally, can move past incremental improvements and deliver fundamentally new innovations in artificial intelligence.

Over the past year, Chinese AI developers such as DeepSeek and other so-called “AI tiger” firms have released a series of advanced large language models that perform competitively with U.S.-based systems on major AI benchmarks. This progress has drawn increasing attention from global investors and technology observers, especially as China intensifies its push to become self-reliant in strategic technologies.

Major Chinese technology companies, including Alibaba, Baidu, and Tencent, have significantly increased their investments in artificial intelligence. These efforts align with Beijing’s broader policy goal of achieving technological independence, particularly in critical areas such as AI infrastructure, data, and computing power. AI has become a national priority, with strong government support encouraging faster development and commercialization.

A crucial component of China’s AI strategy is the development of domestically produced AI chips. While the country has made measurable advances in this area, Hassabis and other industry experts note that Chinese-made chips still lag behind the most advanced processors produced by Nvidia, which remain central to training and deploying state-of-the-art AI models.

Overall, Hassabis’ comments underscore a rapidly evolving global AI race. While the United States and Western firms continue to lead at the cutting edge, China’s AI ecosystem is closing the gap faster than many expected, raising important questions about future competition, innovation, and leadership in artificial intelligence.