China has proposed adding 23 artificial intelligence technologies, from advanced large models to AI agents, to its list of technologies that are banned or restricted from export, The Associated Press reported Friday. The draft signals that Beijing now treats its leading AI capabilities as strategic assets to be controlled, much as Washington has treated advanced chips, and it could affect foreign companies that rely on Chinese models and chip designs.
What Happened
China's Ministry of Commerce and Ministry of Science and Technology jointly published a draft update to the country's catalog of technologies subject to export bans or restrictions, according to the AP report carried by BNN Bloomberg.
The draft covers 40 technologies in all: 23 in AI, 12 in solar power, including methods for making polysilicon, a core raw material for solar panels, and five in biomedicine, including cell and gene therapy. The AP said several chip design technologies, including some for AI processing and memory chips, are also included.
On AI, exports of core technologies for large models trained on sensitive national data would be prohibited, the AP reported. The draft also covers AI agents, software that can carry out tasks on its own, built for highly sensitive settings such as national defense and critical information infrastructure.
Exports of models with 2 trillion or more parameters, or with certain world-leading capabilities, could be restricted. Parameters are the internal values a model learns during training, and their number is a rough gauge of a model's size and capacity. The draft names no models, but the AP noted that Moonshot's Kimi K3 and Alibaba's Qwen 3.8-Max are among Chinese models above that size.
The commerce ministry called the move necessary and cited national security. It said the measures aim to reinforce the country's security baseline, prevent misuse and abuse, and keep AI under human control. The AP report gave no effective date or public comment period.
Why It Matters
For executives, the immediate question is supply. Companies that build products on large Chinese models, or license Chinese AI agents or chip designs, could face new limits or need export approvals once rules are final. The AP account leaves open how "world-leading capabilities" would be judged, which makes the reach of the restriction hard to gauge for now.
The proposal also shows the two largest AI powers converging on the same tool. U.S.-led export controls have kept China from the most advanced AI chips and the machines used to make them, the AP noted. In June, the U.S. Commerce Department asked Anthropic to bar foreign nationals from using its most powerful models, Fable 5 and Mythos 5, citing national security; the AP said those restrictions were later lifted.
Lian Jye Su, chief analyst at research firm Omdia, told the AP the step can be read as a response to Western sovereignty requirements on AI and similar technologies. Gary Ng, a senior economist at Natixis, told the AP that AI has become increasingly strategic for both countries. "The world is getting more fragmented in trade and investment, but also in access to technologies," he said.
Context
The draft arrives as Washington and Beijing try to manage AI risks together. The AP noted progress on AI safety and governance talks and on a mechanism for notifying each other of AI incidents that could threaten national security, part of a package the two governments announced in September.
What to Watch
- When the ministries finalize the catalog, and whether the final text changes the 2-trillion-parameter threshold or defines "world-leading capabilities."
- Whether the rules reach model weights that Chinese labs have already released openly.
- How Chinese AI companies with large models respond, and whether foreign customers face licensing requirements.
- Whether Washington answers with new controls of its own or raises the issue in the bilateral AI talks.
