China’s AI models are rapidly closing in on U.S. capabilities, but breakthroughs remain elusive, according to Demis Hassabis, CEO of Google DeepMind. Hassabis highlighted China’s progress while emphasizing that true frontier innovation has yet to emerge from Chinese firms, speaking on CNBC’s new podcast The Tech Download China’s AI models are rapidly closing in on U.S. capabilities, but major breakthroughs remain elusive, according to Demis Hassabis, CEO of Google DeepMind. Speaking on CNBC’s new podcast The Tech Download, Hassabis noted that Chinese AI is “a matter of months” behind U.S. and Western systems, signaling faster progress than many had anticipated (CNBC, The Tech Download).
Chinese AI Closing In. Hassabis highlighted several Chinese firms making significant strides, including Alibaba, Moonshot AI, Zhipu, and last year’s DeepSeek model, which drew attention for strong performance built on less-advanced chips at lower cost. “Maybe they’re only a matter of months behind at this point,” Hassabis said, emphasizing that while China is catching up, it has yet to demonstrate frontier-level innovation.
Innovation vs. Catch-Up. The DeepMind CEO questioned whether Chinese companies could produce transformative breakthroughs like Google’s 2017 transformer, the scientific advancement that underpins modern large language models such as OpenAI’s ChatGPT and Google’s Gemini. “To invent something is about 100 times harder than it is to copy it. That’s the next frontier really, and I haven’t seen evidence of that yet,” he added.
Challenges and Limitations. China faces hurdles beyond innovation. U.S. export restrictions on high-end Nvidia semiconductors limit access to top-tier hardware for advanced AI development. While Nvidia’s H200 chip may soon be available, domestic alternatives like Huawei’s chips still lag behind. Analysts predict the gap may widen over time as U.S. infrastructure enables faster iteration and more capable AI models. Even Chinese experts acknowledge constraints: Lin Junyang of Alibaba’s Qwen team estimated less than a 20% chance that Chinese firms would surpass U.S. tech giants in AI over the next three to five years.
Hassabis compared DeepMind to a “modern day Bell Labs,” highlighting the importance of exploratory innovation over simply scaling existing knowledge. While China has made rapid progress in AI capabilities, the next breakthrough will require not just technology but the inventive mindset that drives frontier science.