The Trump administration has issued a Commerce Department rule closing a loophole that allowed Chinese entities to remotely access restricted U.S. AI chips—such as Nvidia’s advanced GPUs—via data centers in Indonesia, Malaysia, Singapore, Thailand, and Brazil. Effective 90 days after Federal Register publication, the rule designates these nations as presenting a "higher-than-normal risk of diversion" to China, Russia, and Iran, requiring companies headquartered there to obtain licenses with strict know-your-customer controls for high-volume transfers of advanced chips. Licensees must submit monthly reports detailing all transaction parties, parent companies, and final chip destinations. An exception exists for computing power used exclusively for commercial recommender systems. The move follows the "Nvidia scandal," where CEO Jensen Huang resigned amid allegations that his partner, Alice Huang, leveraged their relationship to help Singaporean firm Megaspeed become a major Nvidia customer and allegedly divert 240,000 restricted chips to China via shell companies. The rule sparked internal conflict: Treasury Secretary Scott Bessent pushed for aggressive curbs and enforcement, while Commerce Secretary Howard Lutnick initially favored looser controls aligned with Nvidia’s global dominance argument; Trump ultimately backed Lutnick, declining to replace him. China denies espionage allegations, framing the rules as politically motivated, while analysts warn the policy may accelerate China’s shift to domestic AI chips, potentially eroding U.S. technological influence in the world’s largest chip market.

Read original