Why it matters: US export controls on advanced AI chips have mostly focused on where the hardware ends up. Now, a new Commerce Department rule dictates who gets to use it. Under the rule, companies must get a license to ship chips from Nvidia and other US suppliers to Malaysia, Thailand, Indonesia, Singapore and Brazil, a move meant to stop Chinese customers from tapping data-center capacity in those countries over the network.
The gap the rule addresses is simple. A Chinese company that can't legally buy a cutting-edge Nvidia processor can still rent time on one, as long as the chip sits in a server rack outside China. The silicon never crosses the Chinese border, but the computing power does.
Firms such as Megaspeed International have built a business on that arrangement. They used billions of dollars' worth of Nvidia chips to build data centers across Southeast Asia, then sold the computing capacity to Chinese tech giants trying to catch up in artificial intelligence. Shutting down that channel is the main goal of the new rule.
"The rule is designed to ensure that existing restrictions on access to sensitive American technology cannot be circumvented through remote computing arrangements involving data centers in third countries," the Commerce Department's Bureau of Industry and Security said in a release.
The rule reaches beyond the five named countries, making it harder for data-center operators anywhere in the world to give Chinese companies remote access to AI chips. It also tightens a separate point in the supply chain by making it more difficult for Chinese chip designers to work with TSMC.
Source link







