Beijing Considers Restricting Overseas Access to China's Top AI Models
SINGAPORE — Chinese authorities have held meetings with top tech firms over the past month about potentially restricting overseas access to China's most advanced AI models, including those yet to be released, three people familiar with the discussions told Reuters.
The Meetings
Companies present at the talks included tech giants Alibaba and ByteDance as well as rising startup Z.ai, whose GLM-5.2 model has recently set Silicon Valley abuzz for matching leading U.S. offerings at a fraction of the cost. The meetings were led by China's Ministry of Commerce.
What's on the Table
Participants discussed putting limits on the most advanced AI models — both closed-source and more open versions. Officials talked about making any leak or theft of proprietary AI technology an offense under China's stringent national security law. They also raised the possibility of implementing new measures to restrict who can fund domestic AI startups.
Some hints of the potential framework emerged from a May roundtable of Chinese legal experts, who proposed a tiered system: basic open-source tools subject to simple filing, more advanced technologies facing security reviews, and the most sensitive frontier models barred from public release or restricted to domestic use.
National Security on Both Sides
The talks follow a number of steps by Beijing to keep homegrown AI within the country and underscore how China, like the U.S., is now treating cutting-edge artificial intelligence as a critical national asset. Since the emergence of DeepSeek's R1 model last year, Chinese AI models have made big inroads globally thanks to their low costs and increasing capabilities.
According to sources, Chinese authorities are also deeply worried about Anthropic's Mythos — the AI model now being used by CISA and the NSA — and its potential to exploit software vulnerabilities against Chinese interests. State media and cybersecurity firm 360 have publicly called for China to develop its own equivalent.
Global Ripple Effects
Any decision by Beijing to limit access to Chinese AI products could ripple across global AI markets. Alibaba's Qwen and ByteDance's Doubao are among the most widely used AI models in China and globally. Costs for many businesses would likely increase if these models become restricted. The scope of potential restrictions is still being discussed and may only apply to future models — but the direction is clear: the era of freely accessible Chinese AI may be coming to an end.
