China’s artificial intelligence models may be just “a matter of months” behind U.S. and Western capabilities, Demis Hassabis, the CEO of Google DeepMind told CNBC.
The assessment from the head of one of the world’s leading AI labs and a key driver behind Google’s Gemini assistant, runs counter to views that have suggested China remains far behind.
Speaking on CNBC’s new podcast, The Tech Download, which launched on Friday, Hassabis said Chinese AI models are closer to U.S. and Western capabilities “than maybe we thought one or two years ago.”
“Maybe they’re only a matter of months behind at this point,” Hassabis told The Tech Download.
About a year ago, Chinese AI lab DeepSeek came out with a model that sent shockwaves through markets because of its strong performance that was built on less-advanced chips and at a lower cost than American alternatives.
While DeepSeek has released new models since, and the shock factor has worn off, China’s tech giants like Alibaba and startups such as Moonshot AI and Zhipu have also released very capable models.
Still, Hassabis said that while China could play catch up, the country’s companies are yet to prove their ability to create AI breakthroughs.
CEO of DeepMind Demis Hassabis listens during a debate at an AI summit at Imperial College London, in central London on July 9, 2025.
Ludovic Marin | Afp | Getty Images
“The question is, can they innovate something new beyond the frontier? So I think they’ve shown they can catch up … and be very close to the frontier … But can they actually innovate something new, like a new transformer … that gets beyond the frontier? I don’t think that’s been shown yet,” Hassabis said.
The transformer was a scientific breakthrough made by Google researchers in 2017 that underpins the large language models that have been developed by AI labs in recent years, including those powering products like OpenAI’s ChatGPT and Google’s Gemini.
Other top technology figures have also given credit to China’s progress. Nvidia CEO Jensen Huang said last year that the U.S. is “not far ahead” in the AI race.
“China is well ahead of us on energy. We are way ahead on chips. They’re right there on infrastructure. They’re right there on AI models,” Huang said.
China chip challenges
China’s technology firms face a number of challenges, with access to critical technology among the biggest. There is a U.S. export ban in force on leading-edge semiconductors from Nvidia that are required to train more advanced AI models.
The White House has indicated that it would approve sales of Nvidia’s H200 chip to China, a more advanced semiconductor than the country recently had access to. However, it is not Nvidia’s top-of-the-range product.
Homegrown chip firms like Huawei have looked to fill the gap, but their performance still lags behind Nvidia’s offering.
Some analysts have suggested that over the longer term, the lack of access to Nvidia chips in China could mean the gap between U.S. and Chinese AI models widens.
“I do suspect, though that we will start seeing a divergence as that superior U.S. AI infrastructure starts iterating those models and starts making those models more capable over time in years to come,” Richard Clode, portfolio manager at Janus Henderson, told CNBC’s “The China Connection” last week.
“So I would expect from here we’re probably at peak relative Chinese AI capability versus the U.S.”
Even Chinese companies have acknowledged their difficulties.
Lin Junyang, technical lead of Alibaba’s Qwen team, said during an AI conference in Beijing last week, that there was a less than 20% chance that a Chinese firm would surpass U.S. tech giants in the next three-to-five years when it comes to AI, the South China Morning Post reported. Lin reportedly said that U.S. computing infrastructure is “one to two orders of magnitude larger” than China’s.
Hassabis however, puts the lack of frontier breakthroughs down to “mentality” rather than tech restrictions.
‘Modern day Bell Labs’
The DeepMind CEO compared the company to a “modern day Bell Labs” which encourages “exploratory innovation” rather than just “scaling out what’s known today. Bell Labs, founded in the early 1900s, was responsible for a number of Nobel Prize-winning discoveries.
“And of course, that’s already very difficult, because you need world-class engineering already to be able to do that. And China definitely has that,” Hassabis said.
“The scientific innovation part that’s a lot harder,” Hassabis added. “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, but it’s very difficult.”
Hassabis is considered to be one of the leading figures in the world of AI. DeepMind, the company he founded more than 10 years ago, which was acquired by Google in 2014, has been a key driving force behind Alphabet-owned Google’s recent success with its AI products, including Gemini.
In November, Google introduced Gemini 3, its latest model, which has been well-received by users and the market as the tech giant looked to allay fears it was falling behind rivals like OpenAI.
