China is quickly catching up to the U.S. in artificial intelligence and may be only months behind

Source Cryptopolitan

Demis Hassabis, the head of DeepMind, says that China is rapidly catching up in the field of artificial intelligence (AI). Hassabis believes that China’s AI may lag only a few months behind the United States. 

China’s big tech giants, including Alibaba and newer startups like Moonshot AI and Zhipu, have developed a range of AI models that have performed extraordinarily effectively over the past few years. One such Chinese firm, DeepSeek, created a model that shocked most experts by performing well even with cheaper computer chips. 

This indicates that the state is not only learning from others’ experience but is also rapidly catching up and closing the gap with leading AI programs worldwide.

Hassabis emphasized that while the country can match much of what the U.S. is doing, the next challenge is innovation—creating something completely new that pushes the world forward. For now, Chinese AI is getting very close to the frontier, but the world will be watching to see if it can go beyond it.

Chinese AI is strong but still behind in original ideas

China, Hassabis said, is well on its way to catching up, but the bigger question is whether it will be able to invent something completely new in AI. He said inventing is about 100 times harder than copying or improving what already exists. 

For instance, he referred to the Transformer, an idea first proposed by Google researchers in 2017. That notion reshaped how artificial intelligence understands language, solves problems, and generates answers. Today, it is the backbone of a wide range of modern AI systems, including ChatGPT and Google Gemini. Hassabis observed that, until now, Chinese companies had not developed anything new that could completely revolutionize AI.

He compared DeepMind to Bell Labs, a famous research center from the past that made many major discoveries. Bell Labs encouraged scientists to explore and try new ideas.

Hassabis said DeepMind works the same way, trying new ideas rather than just copying. He believes the next big step in AI is innovation — coming up with completely new ideas — and that isn’t easy, even for China.

Chip shortages slow China’s AI progress

A challenge China is encountering is the shortage of computer chips. Strong AI requires very strong computers, and top-tier ones use chips from companies such as Nvidia in the United States. Currently, China cannot purchase the very best chips due to U.S. government regulations. It may be allowed to sell some advanced chips, but not the top ones. 

Chinese companies such as Huawei attempt to develop their own chips, but these are not as strong as Nvidia’s best. That means the state might struggle to train the biggest and most powerful AI programs. And over time, some experts say, this could widen the gap between American and Chinese AI.

Even leaders in China agree it is difficult. A technical expert from Alibaba recently said there is less than a 20% chance that a Chinese company will surpass the U.S. in AI over the next three to five years. He said the U.S. has much more powerful computer resources, which makes a big difference.

Still, Hassabis thinks the challenge is not just technology — it is also about how people think and invent. China has brilliant engineers, but creating something completely new is still very hard.

Other big tech leaders also notice China’s progress. Jensen Huang, the boss of Nvidia, said that China is strong in some areas, such as energy and infrastructure. He said the U.S. is ahead in chips, but China is close in AI models.

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.

Disclaimer: For information purposes only. Past performance is not indicative of future results.
placeholder
Bitcoin shows strong correlation with institutional demand following 7% uptickBitcoin's price has largely tracked net institutional demand over the past year, according to Bitwise. Net institutional demand is the buying activity of global exchange-traded products (ETPs) and treasury companies minus new supply.
Author  FXStreet
Yesterday 01: 37
Bitcoin's price has largely tracked net institutional demand over the past year, according to Bitwise. Net institutional demand is the buying activity of global exchange-traded products (ETPs) and treasury companies minus new supply.
placeholder
Silver Price Forecast: XAG/USD corrects to near $86.50 as Iran stops killing protestersSilver price corrects almost 6% to near $86.50 during the Asian trading session on Thursday.
Author  FXStreet
Yesterday 05: 55
Silver price corrects almost 6% to near $86.50 during the Asian trading session on Thursday.
placeholder
Standard Chartered lifts Ethereum call to $7,500, arguing institutional demand could leave Bitcoin trailingStandard Chartered raised its year-end Ethereum target to $7,500 (from $4,000), citing institutional demand, while projecting $25,000 by 2028 and scenarios toward $40,000 by 2030 amid ETF- and treasury-driven accumulation.
Author  Mitrade
Yesterday 06: 11
Standard Chartered raised its year-end Ethereum target to $7,500 (from $4,000), citing institutional demand, while projecting $25,000 by 2028 and scenarios toward $40,000 by 2030 amid ETF- and treasury-driven accumulation.
placeholder
Bitcoin Hits $97K Despite Elevated PPI and Lack of US Tariff DecisionDespite higher-than-expected Producer Price Index (PPI) inflation data for November, Bitcoin surged to new eight-week highs, diverging from US stock markets.
Author  Mitrade
20 hours ago
Despite higher-than-expected Producer Price Index (PPI) inflation data for November, Bitcoin surged to new eight-week highs, diverging from US stock markets.
placeholder
Bitcoin Flashes Classic Bottom Signals as BTC Nears $101K ReclaimBitcoin nears two-month highs with key indicators signaling potential for further gains as it targets $101,000.
Author  Mitrade
2 hours ago
Bitcoin nears two-month highs with key indicators signaling potential for further gains as it targets $101,000.
goTop
quote