The US Department of Commerce just plugged a hole in its semiconductor export controls that was roughly the size of a shipping container full of advanced AI chips. Updated guidance issued on May 31 now requires export licenses for advanced computing items sold to entities headquartered in China, even when those entities operate through subsidiaries in countries like Malaysia and Singapore.
The loophole, which reportedly allowed hundreds of thousands of Nvidia’s Blackwell and Rubin processors to reach Chinese-controlled operations through indirect channels, had been an open secret in the chip industry.
What the new rules actually change
Previously, Chinese companies could set up shop in friendly third countries and purchase advanced Nvidia chips without triggering the export license requirements that would apply to a direct sale to China. The Commerce Department’s updated guidance targets entities headquartered in Country Group D:5 (which includes China) or Macau, regardless of where the actual purchasing subsidiary sits.
The restrictions have been evolving since 2022, with various policy adjustments along the way. A December 2025 allowance permitted limited exports of the H200 variant, and continued revisions throughout early 2026 have kept the industry guessing about the next move. The AI Diffusion rule implemented in 2025 was supposed to be the comprehensive framework, but the subsidiary workaround became one of the most significant gaps in the entire export control regime.
Nvidia CEO Jensen Huang has already acknowledged the reality on the ground. He stated that the company has “largely conceded” much of the direct AI chip market in China to domestic competitors like Huawei.
The competitive landscape shifts
For Nvidia investors, the loophole closure removes what was effectively a backdoor revenue stream that had been partially cushioning the blow of formal restrictions. Given that the subsidiary workaround was facilitating purchases of hundreds of thousands of advanced chips, the scale of the now-blocked demand suggests the market may not have fully accounted for this particular channel drying up.
Why crypto markets should pay attention
Projects building decentralized AI compute networks, tokenized GPU marketplaces, and AI-powered DeFi protocols all have some degree of dependency on the availability and pricing of advanced chips. AI-focused tokens have already shown sensitivity to Nvidia-related news cycles, and tokenized equity products linked to Nvidia or the broader semiconductor sector face similar exposure.