China's Ministry of Commerce is actively discussing with leading tech giants — Alibaba, ByteDance, and Z.ai — the possibility of imposing restrictions on foreign users' access to China's most advanced artificial intelligence models. This information comes from reliable insider sources, and I am inclined to view it as a signal of a fundamental shift in Beijing's strategy.
The proposed measures will likely only affect future generations of models, not those already released to the market. The key idea is to equate the leakage or theft of proprietary AI technologies with a violation of national security law. Separately, tightening control over foreign fund investments in Chinese AI startups is being discussed.
This decision is not just protectionism. It is a direct response to concerns that American cyber models, such as Mythos, could be used to find vulnerabilities in Chinese infrastructure. Beijing is making it clear: technological sovereignty now outweighs global expansion.
It is important to understand that such steps risk undermining China's main competitive advantage in the global AI market. In 2025–2026, Chinese models actively gained popularity abroad precisely due to low cost, open weights, and a liberal access regime. Restrictions could weaken this channel, which allowed China to effectively compete with American developments.
Preventive Measures and Escalation of Conflict
Beijing's actions are not a spontaneous reaction. Previously, China blocked Meta's deal to acquire startup Manus for $2 billion, and in June tightened control over cross-border transactions involving Chinese technologies. Simultaneously, an expert discussion at China's Supreme People's Court considered a differentiated approach to open source: from simple registration of basic tools to a complete ban on public releases for sensitive models.
This occurs against the backdrop of the US expanding export controls from chips to access to AI models themselves. On June 12, Anthropic, at the request of US authorities, restricted access to its Fable 5 and Mythos 5 models for foreign citizens. Although the directive was later revoked, the incident involving the distillation of Claude, for which operators linked to Alibaba were blamed, only added fuel to the fire.
Alibaba has already announced a ban on the use of Anthropic products for employees starting July 10, suggesting a switch to its own Qoder platform. Meanwhile, 360 Security Technology founder Zhou Hongyi introduced the Tulong Feng tool as a direct Chinese response to Mythos 5.
Cryptalist Analysis: We are witnessing a transition from a trade war to a war of algorithms. Both sides understand that control over advanced AI models is not a matter of commerce, but of national security. China, having recognized its vulnerability in the "chip war," is now building a "digital wall" around its developments. On one hand, this will protect technologies, but on the other, it may isolate Chinese developers from the global market and slow their long-term progress. The market should prepare for the fragmentation of the AI ecosystem into two independent poles.