New US rules aim to restrict Chinese firms’ remote access to AI servers and infrastructure

New steps have been taken to ensure that countries and companies that are not on whitelists that would allow them to import advanced AI hardware cannot use cloud-based workarounds. The U.S. Commerce Department’s Bureau of Industry and Security (BIS) has today unveiled an interim final rule, or IFR, in collaboration with the Department of Treasury and the National Security Council, aimed at backing up existing limitations on the export of hardware built by companies like Nvidia to non-whitelisted companies based in countries such as China. "The rule is designed to ensure that existing restrictions on access to sensitive American technology cannot be circumvented through remote computing arrangements" using data centers in third-party countries, a statement explains.

- While the U.S. flip-flops on chip sanctions, China is building its own chip supply market
- The state of China's decade-long semiconductor push
- Huawei Ascend NPU roadmap examined
- China's CXMT targets 30% DRAM memory market share by 2030 with sixth mega-fab
- China's latest round of rare-earth export controls explained
Nvidia, for example, is only allowed to export its high-performance AI chips to companies that are on a whitelist. But amid ongoing claims of black market GPU sales and a recent crackdown on overseas AI GPU rentals from China, it has become clear that these restrictions perhaps aren't as robust as they were intended to be. Nvidia Blackwell Ultra B300 GPUs were even recently discovered in a data center used by China Telecom.
But while blocking the sale of physical goods is one thing, the cloud-based workaround is another entirely. Nvidia itself told data center operators earlier this month that they must limit remote access to their services — 51% of their business must come from non-Chinese companies. But, again, it's unclear how effective that approach has been in ensuring access to powerful AI capabilities doesn't fall into the wrong hands, even if they're accessed remotely rather than in person.
The new BIS rule means that overseas data centers will be required to do more to limit access to their services where U.S-based technology is being used. Restrictions will also be placed on data centers that supply remote access to companies whose headquarters are deemed to be in higher-risk countries, such as China, Russia, and Iran. A license will also be required for anyone trying to export advanced chips to companies with headquarters in Indonesia, Malaysia, Singapore, Thailand, or Brazil. The new requirements will officially go into effect in 90 days.
This latest rule comes against a backdrop of ongoing uncertainty over America's AI crown jewel, Nvidia. Jensen Huang, the company's CEO, resigned after confirming an affair, but is now potentially in line for a quick-fire comeback, despite the fact that the woman involved in the affair is an alleged Chinese spy.
