Chinese state-linked entities are leveraging cloud services provided by Amazon and other U.S. tech giants to gain access to advanced AI chips and technologies that are otherwise restricted due to U.S. export controls. These entities are utilizing cloud platforms to tap into high-performance computing power and generative AI models, circumventing direct acquisitions of restricted hardware and software. The strategy has raised concerns among U.S. lawmakers and regulators, who are now considering tighter controls on such indirect access.
The U.S. government has imposed restrictions on the export of high-end AI chips to China, aiming to limit the enhancement of the Chinese military’s capabilities. However, the use of cloud services to access these technologies does not violate current regulations, as the export controls only apply to direct transfers of physical goods or software. Public tender documents reveal that Chinese institutions, including universities and research labs, have been acquiring cloud-based AI capabilities through intermediaries.
For instance, Shenzhen University purchased access to cloud servers powered by Nvidia A100 and H100 chips, which are restricted for export to China. The university obtained these services through a Chinese intermediary, Yunda Technology, to support an unspecified project. Similarly, Zhejiang Lab, a research institute, intended to use AWS cloud services for its AI model, GeoGPT, due to insufficient computing power from domestic providers like Alibaba Cloud.
The U.S. government is now exploring measures to close this regulatory gap. Proposed legislation would empower the Commerce Department to regulate remote access to U.S. technology via cloud services. Additionally, a proposed rule could require U.S. cloud providers to verify the identity of large AI model users and report their activities to regulators.
This loophole has been a growing concern for U.S. officials, including Michael McCaul, chair of the U.S. House of Representatives Foreign Affairs Committee. The Commerce Department has acknowledged the issue and is working with Congress to strengthen existing controls, potentially restricting Chinese companies’ access to advanced AI capabilities through cloud computing.
Amazon has stated that it complies with all applicable U.S. laws regarding its cloud services, both domestically and internationally. However, the company recently updated its Chinese-language promotional materials to clarify that some services are not available in China. Microsoft’s cloud services have also been targeted by Chinese entities, as evidenced by Sichuan University’s purchase of Microsoft Azure OpenAI tokens to support a generative AI project.
These developments highlight the complex interplay between global demand for AI technology and the regulatory frameworks designed to control its distribution. As Chinese entities continue to seek access to advanced computing resources, U.S. regulators are under pressure to adapt and enforce controls that address the evolving landscape of cloud-based technology transfers.
READ MORE: