U.S. tightens controls on China’s overseas GPU rentals
The U.S. Commerce Department’s Bureau of Industry and Security (BIS) has issued an interim final rule targeting remote access to advanced AI chips and servers. The measure addresses a gap that allowed Chinese companies barred from obtaining NVIDIA’s latest GPUs to rent computing power overseas instead. BIS says the rule was written by and at the direction of the Department of the Treasury and the National Security Council.
A license is required to ship advanced accelerators to any data center worldwide providing more than 12 million ‘aggregate TPP’ of computing power to companies headquartered in Macau or Country Group D:5, which includes China. TPP, or Total Processing Performance, measures chip computing power using peak dense throughput multiplied by the operation’s bit length. The aggregate measure counts every accelerator in the facility.
The restrictions also cover linked data centers and indirect access, including cloud resellers training models for Chinese firms. Facilities already serving affected customers have 90 days to obtain a license. Only commercial recommender systems are exempt; systems generating text, images, video, or code do not qualify.