（Singapore, 07.07.2026）China is considering restrictions on overseas access to its most advanced artificial intelligence models as Beijing steps up efforts to protect homegrown technology and treat cutting-edge AI as a critical national asset.
According to Reuters, Chinese authorities have held meetings with leading technology companies over the past month to discuss possible controls on advanced AI models, including products that have yet to be released.
Companies involved in the discussions included Alibaba, ByteDance and Chinese AI startup Z.ai, according to people familiar with the matter. The talks were led by China’s Ministry of Commerce, while officials from the National Development and Reform Commission also attended.
The discussions come as Chinese AI models gain popularity around the world due to their improving capabilities and relatively low costs. The rapid rise of DeepSeek’s R1 model last year helped increase global interest in Chinese AI technology and encouraged businesses and developers to explore alternatives to more expensive U.S. models.
Any move by Beijing to restrict overseas access could have a significant impact on the global AI industry, potentially reducing the availability of lower-cost models and raising expenses for businesses that rely on Chinese technology.
People familiar with the discussions said officials considered imposing controls on the country’s most advanced AI systems, including both closed-source and open-weight models. Open-weight models allow users to download, run and customize the underlying systems.
The scope of any potential restrictions is still under discussion and may only apply to future models. It remains unclear when, or whether, the proposed measures will be introduced.
Officials also discussed tougher penalties for the leak or theft of proprietary AI technology, including the possibility of treating such cases as offences under China’s national security laws.
Authorities are also considering new restrictions on who can invest in domestic AI startups, according to Reuters.
China’s Ministry of Commerce, the National Development and Reform Commission and the companies involved in the meetings did not respond to requests for comment.
Alibaba’s Qwen and ByteDance’s Doubao are among the most widely used AI models in China. Z.ai has also attracted attention in Silicon Valley after its GLM-5.2 model demonstrated capabilities close to leading U.S. systems at a significantly lower cost.
The latest discussions highlight growing concerns in both China and the United States over the national security implications of increasingly powerful AI technology.
The U.S. government has also taken steps to restrict access to advanced American AI systems over concerns that they could be used by military or intelligence organizations in countries including China and Russia.
In June, the Trump administration ordered restrictions on foreign access to Anthropic’s most advanced Fable and Mythos models. The company initially disabled the models globally because users’ nationalities could not be verified in real time.
Restrictions on Fable, a model designed for general users, were later lifted after additional safeguards were introduced. Mythos, which was developed for cybersecurity professionals, remains available only to selected trusted U.S. organisations.
Some U.S. AI experts have also called for tighter regulation of Chinese AI models.
Chinese authorities, meanwhile, are reportedly concerned about the potential security risks posed by Mythos, particularly its ability to identify and exploit software vulnerabilities.
Those concerns have also been raised publicly by Chinese state media and Zhou Hongyi, founder of cybersecurity company 360, who has called on China to develop its own alternative to Mythos.
Beijing has introduced several measures this year to strengthen control over domestic AI technology and overseas investments.
In April, Chinese authorities ordered Meta to unwind its US$2 billion acquisition of Chinese-founded AI startup Manus. New regulations introduced in June also tightened oversight of overseas transactions involving Chinese investors, technology, data and national security.
Authorities have also investigated Manus and other Chinese AI startups that moved operations overseas to determine whether they violated export control laws, according to Reuters.
It remains unclear how restrictions on overseas access to Chinese AI models would be implemented.
However, a discussion among Chinese legal experts in May offered some indication of how future regulations could work.
According to a summary published in an official Supreme People’s Court journal, experts proposed a tiered regulatory system for open-source AI technology.
Under the proposal, basic open-source tools would only require simple registration, while more advanced technologies would undergo security reviews. The most sensitive frontier AI models could be prohibited from public release or restricted to domestic use.
The discussions come as competition between China and the United States increasingly expands beyond semiconductor chips and computing infrastructure to include access to advanced AI models.
As Chinese AI companies gain a larger global following, potential restrictions could reshape the international AI market and determine how widely the country’s next generation of technology can be used outside China.
