China Proposes New Export Controls on 23 AI Technologies

This proposal reflects China's strategic pivot in the global tech landscape amid rising geopolitical tensions.

• On October 16, 2026, China proposed adding 23 AI technologies to its export control catalogue.
• The draft targets large models, AI agents, and computing clusters, with specific thresholds for regulation.
• Public consultation is open until November 15, 2026, with immediate registration required for affected exports.

• Geopolitical competition: This move comes as the U.S. imposes restrictions on advanced AI technologies, prompting China to bolster its domestic capabilities.
• National security focus: The proposal aims to prevent misuse of AI technologies and ensure human oversight amid rapid advancements.
• Market implications: Companies may pivot towards domestic alternatives or accelerate open-source strategies in response to these controls.

On October 16, 2026, China's Ministry of Commerce and Ministry of Science and Technology released a draft proposal to expand the Catalogue of Technologies Prohibited and Restricted from Export. This catalogue now includes 23 AI technologies, bringing the total to 40. The targeted technologies encompass large AI models, AI agents, and computing clusters, with specific controls triggered by training on sensitive national data or exceeding defined thresholds—such as 2 trillion parameters or 10^26 FLOPs of compute power.

The proposal distinguishes between prohibited items, like autonomous malware tools, and restricted items, such as large model weights that surpass specified limits. Notably, the draft includes a fast-track automatic licensing process for open-source releases and certain overseas activities by Chinese firms, which could facilitate international collaboration while maintaining regulatory oversight.

This regulatory shift is part of China's broader strategy to enhance national security and prevent the misuse of advanced technologies. The Chinese government is keenly aware of the implications of AI on national defense and public safety, especially in the context of ongoing U.S.-China technology competition. By tightening export controls, China aims to ensure that sensitive technologies do not fall into the hands of adversaries, while also fostering a domestic ecosystem capable of innovation.

International observers have drawn parallels between these proposed controls and existing U.S. regulations on model weights, suggesting a tit-for-tat dynamic in global tech governance. As companies navigate these new restrictions, they may increasingly turn to domestic alternatives or accelerate their open-source initiatives to mitigate the impact of export controls.

The public consultation period, which runs until November 15, 2026, allows stakeholders to voice their opinions on the proposed measures. However, the immediate registration requirement for affected exports indicates that the Chinese government is moving quickly to implement these controls, reflecting a sense of urgency in addressing national security concerns.

• Tech companies: Firms relying on Chinese AI technologies may face immediate licensing challenges.
• Developers: Those working with large AI models could see restrictions on their projects.
• Businesses in Dubai: Local enterprises may experience indirect effects through altered availability of certain AI models.

• Public feedback: Monitor the outcomes of the public consultation period and any adjustments to the proposal based on stakeholder input.
• International responses: Watch for reactions from other nations, particularly the U.S., which may influence global tech policies.
• Market shifts: Keep an eye on how companies adapt, particularly in terms of sourcing AI technologies and pivoting to domestic alternatives.