The U.S. is trying to stop China from renting AI compute abroad, but China has already built a parallel supply chain using U.S. AI outputs
The US wants to block China from renting AI power abroad but China already built a parallel supply chain from US AI output
The U.S. Department of Commerce's Bureau of Industry and Security (BIS) issued an interim rule requiring data center operators in Singapore, Malaysia, Thailand, Indonesia and Brazil to obtain U.S. licenses to buy advanced U.S.-made AI chips such as Nvidia's. The rule brings remote access under controls consistent with physical export bans, and takes effect upon publication in the Federal Register; Nvidia shares fell as much as 2.9%, and the Philadelphia Semiconductor Index fell as much as 5.5%. The trigger for the new rule was suspected chip smuggling; Nvidia CEO Jensen Huang suddenly resigned recently amid an extramarital affair scandal, and U.S. authorities are investigating whether Singapore's Megaspeed violated export controls, which Megaspeed denies. Separately, according to a Reuters investigation after reviewing more than 300 Chinese papers, patents, job postings and government documents, China has turned model distillation into an industry for coping with shortages of high-end compute, using software to automatically extract answers and code from top U.S. models to train smaller local models; distillation is not illegal but violates OpenAI's and Anthropic's terms of service, and Anthropic says it can already block most such attacks.
The article places the new U.S. rules requiring licenses for U.S.-made AI chips destined for overseas data centers alongside Reuters' investigation of China's distillation industry chain, presenting the interaction between controls and alternative pathways.