Beijing is preparing a major shift in its AI policy. In recent weeks, China's Ministry of Commerce has held consultations with leading tech giants — Alibaba, ByteDance, and Z.ai — on introducing restrictions on foreign users' access to the most advanced Chinese artificial intelligence models. This is not just a bureaucratic initiative, but a signal of the beginning of an era of technological nationalism in the AI sector.
What exactly is being proposed to change?
According to data obtained from insider circles, the restrictions will only affect future models, not those already released on the market. Officials are considering recognizing the leakage or theft of proprietary AI technologies as a violation of national security law. Separately, tighter control over foreign investments in Chinese AI startups is being discussed.
This move is directly linked to Beijing's growing anxiety about American cyber models. There are concerns that systems such as Mythos could be used to find and exploit vulnerabilities in China's critical infrastructure in Washington's interests. Thus, China is trying not only to protect its technologies but also to prevent their use against itself.
The price of sovereignty
Such decisions carry serious risks. Restricting access to Chinese models could weaken the main channel through which China competes with American developments. In 2025–2026, Chinese models actively captured the global market due to low cost, open weights, and a less stringent access regime. Introducing barriers would jeopardize this key advantage.
These discussions are taking place against the backdrop of tightening export controls by the United States, which have moved from restrictions on chips to controlling access to AI models themselves. On June 12, Anthropic, at the request of U.S. authorities, restricted access to Fable 5 and Mythos 5 for foreign nationals, including its own employees outside the country. This created a precedent that now appears to be followed by Beijing as well.
China has already taken concrete steps earlier: in April, the National Development and Reform Commission canceled Meta's deal to acquire the startup Manus for $2 billion, and in June, control over foreign transactions involving Chinese investors was tightened. To complete the picture, in May, an expert discussion was held at the Supreme People's Court, where a differentiated approach to open source was proposed: from simple registration for basic tools to a complete ban on public release for sensitive models.
Expert commentary: China is consciously risking the loss of the global market to secure its technological sovereignty. However, in the long term, this could lead to the fragmentation of the AI ecosystem, where each major power develops its own "closed" models. For investors and developers, this means a fundamental change in the rules of the game: the openness and accessibility that drove progress in AI may give way to fierce geopolitical competition.