China moves to restrict global access to its most advanced AI models

Chinese authorities have held meetings with top technology companies over the past month to discuss potentially restricting overseas access to the country’s most advanced artificial intelligence models, including those yet to be released.

According to Reuters, citing people familiar with the development, the move signals that Beijing is treating cutting-edge AI as a critical national security asset.

Three people familiar with the discussions, who were not authorised to speak to media and declined to be identified, said the talks were led by China’s Ministry of Commerce and included tech giants Alibaba and ByteDance as well as AI startup Z.ai.

The discussions follow a series of steps by Beijing to keep homegrown AI within the country, and mirror similar moves by the United States to restrict the export of advanced AI technology to foreign markets.

What they are saying

According to Reuters, two of the sources said participants at the meetings discussed putting limits on the most advanced AI models, covering both closed-source and more open versions of the technology.

• Officials raised the possibility of making any leak or theft of proprietary AI technology an offence under China’s stringent national security law, one of the sources said.
• This move would significantly raise the legal stakes around the handling of advanced model weights and training data.
• Authorities also discussed implementing new measures to restrict who can fund domestic AI startups, potentially limiting the flow of foreign capital into China’s AI sector, according to the same source cited by Reuters.
• The scope of the potential restrictions is still being debated, with two sources noting that the measures may apply only to future models rather than existing ones.

It was not immediately clear when or even whether the restrictions would come into force.

More insights

The talks come against the backdrop of Chinese AI models making significant global inroads over the past year, driven by their low costs and rapidly improving capabilities following the emergence of DeepSeek’s R1 model in 2024.

Any decision by Beijing to limit overseas access to its most capable AI models could ripple across global AI markets, as many businesses outside China have increasingly relied on Chinese models as cost-effective alternatives to US-developed systems from companies such as OpenAI and Anthropic.

The potential restrictions would affect both companies that access Chinese models through APIs and those that deploy open-weight versions locally, given that the discussions reportedly cover both closed-source and open model categories.

What you should know

China’s move to potentially restrict AI model exports mirrors a broader global pattern in which governments are increasingly treating advanced AI as a strategic national asset.

• In mid-June, Nairametrics reported that AI gaint, Anthropic shut down access to its two most advanced artificial intelligence models, Fable 5 and Mythos 5, for users globally following a US government export control directive linked to national security concerns.
• The company said it received the directive requiring it to block access to both models for all foreign nationals, whether they are located inside or outside the United States. The restriction also applies to Anthropic’s own employees who are foreign nationals.
• The US government eventually eased the ban. On June 27, Nairametrics reported that the US government had partially lifted the two-week-old suspension of access to Anthropic’s most powerful AI models.

This partial reversal allowed more than 100 companies and institutions, including many Fortune 500 companies, to regain access to Claude Mythos 5, though the model remains restricted from general public use.