Key Points
The U.S. requires licenses for advanced AI chip exports to Singapore, Malaysia, Thailand, Indonesia and Brazil, four of which are in Southeast Asia. A separate control requires a license for chips in any data center that provides firms headquartered in China or other restricted destinations with more than 12 million TPP, equal to about 760 H100s or 200 B300s.
Non-U.S. cloud and GPU rental operators serving mainly Chinese customers face the most direct impact, while U.S. CSPs are exempt from the five-country licenses, provided exporters report monthly.
Chinese CSPs may turn to domestic AI chips, in-house ASICs and compliant NVIDIA chips such as the RTX PRO 5500, though supply bottlenecks could limit availability.
Foundries and OSATs need licenses to ship advanced logic chips for non-approved designers located in, or with an ultimate parent in, the five countries or restricted destinations, and must screen both designers and end users.
Related report: US Remote-Access Rule Reshapes AI Compute in SEA


