China is set to implement new export restrictions on artificial intelligence (AI) technology that will enable foreign customers to access Chinese AI services while allowing the Chinese government to maintain tighter control over essential models and infrastructure. The draft regulations, released on October 16, reflect a strategy aimed at balancing international access to Chinese AI with safeguarding domestic tech assets.
The proposed rules establish a clear distinction between foreign access to AI capabilities through platforms and websites and the acquisition of the underlying technology necessary for independent operation or development of similar systems. Specifically, the regulations will add 23 AI-related technologies to a list of those banned or restricted from export, including advanced AI models and technologies for designing semiconductors.
These proposals are anticipated to be adopted by the end of 2026, with the draft currently open for public comments until November 15. A source familiar with the discussions noted that the final implementation guidelines will follow this consultation period.
This framework prioritizes a business model where intellectual property and research remain in China while international users pay for access to AI services. Alfredo Montufar-Helu, managing director at the consultancy Ankura, mentioned that this approach aligns with existing market trends aimed at retaining domestic control over R&D while monetizing services abroad.
Companies like DeepSeek, Moonshot, and Z.AI already provide AI services on a global scale through cloud platforms, and such arrangements are expected to continue under the new regulations. However, certain international deployments and technology transfers will require government approval.
The proposed regulations also address security concerns, as advanced AI systems may support sensitive data processing and cyber operations. By maintaining control over these technologies, Chinese authorities can better oversee how these capabilities are deployed and utilized outside China.
China’s AI developers face challenges amid U.S. export controls limiting access to advanced computing hardware necessary for training AI models. Washington’s restrictions on certain high-performance processors have led some companies to look for overseas data centers to obtain the required computing power. Yet, the new Chinese rules may complicate that strategy as researchers seek clarification on operating their advanced models on foreign servers.
Legal experts have expressed concerns regarding the imprecise criteria for categorizing AI technologies under the new export controls. These include thresholds for computational capacity and model size, leaving companies uncertain about which models might fall under the restrictions. Companies are also grappling with how transactions, especially those involving foreign investment or acquisitions of Chinese AI technology, could be affected by the new regulations.
The strategic moves from both Beijing and Washington illustrate a competitive landscape where technology is tightly regulated. As both nations navigate this evolving terrain, the outcome will likely shape the future of AI development and international collaboration in the technology sector.
The upcoming November 15 deadline for public comment will be crucial in determining the final parameters of China’s export controls, which could have significant implications for AI developers and international investors alike.
