Demis Hassabis, CEO of Google’s artificial intelligence (AI) company DeepMind, believes that Chinese AI firms have yet to achieve breakthrough innovations at the cutting edge of technology, and overall remain about six months behind the most advanced AI labs in the West.
According to Bloomberg, Hassabis said in an interview at the World Economic Forum in Davos, Switzerland, that when Chinese AI company DeepSeek released its R1 model a year ago, the reaction from the outside world was "a massive overreaction."
Hassabis said: “They are very good at catching up with the technological frontier, and this capability is getting stronger. But I think they have not yet demonstrated the ability to innovate beyond the frontier.”
Currently, Chinese AI companies face multiple restrictions. The United States is preventing these firms from obtaining the most advanced semiconductors needed to develop and operate AI, forcing Chinese researchers to use unconventional methods and architectures.
However, some restrictions are expected to be eased. U.S. President Trump is relaxing the ban on exporting advanced AI chips to China, including conditionally approving American tech giant Nvidia to export H200 chips to China, marking a significant shift in U.S. policy.
Nonetheless, for national security reasons, sales of the most advanced AI processors will still be restricted, and this move remains highly controversial. DeepMind competitor Anthropic’s CEO, Amodei, pointed out on Tuesday that exporting Nvidia chips to China is “like selling nuclear weapons to North Korea.”