Key Takeaways:
The U.S. Commerce Department on Sunday closed a year-old loophole that may have allowed hundreds of thousands of advanced AI chips from Nvidia Corp. and Advanced Micro Devices Inc. to reach Chinese companies through overseas subsidiaries, escalating Washington's campaign to restrict Beijing's access to cutting-edge semiconductor technology.
"This is a HUGE problem," Chris McGuire, a former State Department official and technology policy expert, said in a social media post. "Chinese companies have been buying these chips, very likely at scale." McGuire said the loophole allowed overseas subsidiaries of Chinese firms to purchase Nvidia's most advanced Blackwell processors without an export license.
The guidance, posted on the Commerce Department's website Sunday, closes a gap created when the Trump administration paused enforcement of the Biden-era AI Diffusion Rule in May 2025. That rule had governed global access to advanced AI chips but was not actively enforced for roughly a year, creating what officials now view as an unintended opening. One chip industry source with deep supply-chain knowledge estimated that hundreds of thousands of processors — including Nvidia's Rubin and Blackwell architectures and AMD's MI350x series — may have moved through indirect channels to Chinese subsidiaries in locations such as Malaysia.
The updated guidance shifts enforcement from physical shipment geography to corporate ownership, requiring export licenses for entities headquartered in China even when operating through subsidiaries in third countries. The Commerce Department said it would enforce license requirements for advanced chips to entities headquartered in China regardless of where those entities are physically located. The affected chips represent the frontier of AI computing — Nvidia's Blackwell and Rubin processors power the largest large language models, while AMD's MI350x competes directly in the data center GPU market.
The regulatory tightening comes as the U.S. semiconductor export control regime faces growing enforcement challenges. The previous average U.S. tariff on Chinese goods stands at roughly 19 percent after multiple escalation rounds since 2018, according to the Peterson Institute for International Economics, and the chip-specific controls represent a separate, more targeted layer of technology containment. After the initial October 2022 export controls on advanced semiconductors, Nvidia's data center revenue in China fell sharply while the company redirected supply to other markets.
The new guidance does not require data centers to stop using chips already acquired or to cut off servicing of advanced computing systems already deployed. This suggests Washington is prioritizing forward-looking containment over retroactive enforcement, though the scale of potential diversion — estimated in the hundreds of thousands of units — raises questions about how effectively the earlier policy gap was monitored.
For Nvidia and AMD, the implications are significant. Nvidia's data center segment generated $47.5 billion in revenue in its most recent fiscal year, with China historically accounting for roughly 15 to 20 percent of that total before the initial export restrictions took effect. The closure of this loophole eliminates a channel that may have partially offset those earlier restrictions. Both companies declined to comment on the guidance.
The action signals a broader tightening of U.S. controls over AI infrastructure, expanding scrutiny from physical shipment destinations to corporate structures and indirect procurement pathways. With the U.S. presidential election cycle approaching, technology export policy toward China is likely to remain a central axis of geopolitical competition, and further regulatory refinements targeting indirect access channels are expected in coming months.
This article is for informational purposes only and does not constitute investment advice.
