China just 'months behind' US in AI race: DeepMind CEO
What's the story
Demis Hassabis, the CEO of Google DeepMind, has warned that China is closing the gap with the US in artificial intelligence (AI) at an unprecedented pace. He made the remarks on CNBC's podcast The Tech Download, which was released on Friday. The statement challenges the widely held belief that China is still years away from matching US capabilities in AI technology.
Rapid progress
China's AI models are just months behind US
Hassabis said, "Maybe they're only a matter of months behind at this point," emphasizing that Chinese models are much closer to the AI frontier than previously thought. The statement highlights China's rapid progress in the field over the last couple of years, narrowing the gap with US tech much faster than expected.
Tech breakthrough
DeepSeek's powerful model and China's AI advancements
The rapid advancement of Chinese AI was evident last year when AI lab DeepSeek unveiled a powerful model that shocked the global tech industry. The system performed well on less-advanced chips and was cheaper to build than its American counterparts. Since then, Chinese tech giants like Alibaba and start-ups such as Moonshot AI and Zhipu have been churning out more advanced models.
Tech innovation
Innovation vs. imitation: Hassabis's perspective on AI race
Despite China's rapid progress, Hassabis believes there's a key difference between catching up and leading in the AI race. He questioned if Chinese firms can innovate something new beyond the frontier, citing Google's 2017 transformer architecture as an example of a game-changing breakthrough. This innovation now powers systems such as OpenAI's ChatGPT and Google's Gemini.
Industry recognition
NVIDIA CEO acknowledges China's AI progress
NVIDIA CEO Jensen Huang has also acknowledged China's rapid progress in the AI race. He said that while the US is ahead, China is not far behind in infrastructure and model development. However, access to advanced semiconductors remains a major hurdle for Chinese firms due to US export restrictions. Despite domestic chipmakers like Huawei trying to bridge this gap, their technology still lags behind NVIDIA's top offerings.
Mindset matters
Hassabis emphasizes mindset over technology in AI race
Hassabis believes the real challenge in the AI race isn't technology, but rather mindset. He compared DeepMind to a "modern-day Bell Labs," emphasizing deep scientific innovation over merely scaling existing technologies. "To invent something new is about 100 times harder than copying it," he said, stressing that this remains the true next frontier in AI development.