The United States government has officially confirmed that its licensing requirements for advanced AI semiconductor exports extend to overseas subsidiaries of Chinese companies — not just entities operating within China’s borders. The Department of Commerce issued this formal guidance on Sunday, June 1, 2026, responding to widespread concerns that existing restrictions had a significant enforcement gap.
The clarification came from the Bureau of Industry and Security (BIS), the Commerce Department division responsible for administering export controls. BIS stated it was responding to direct inquiries about whether prior licensing rules were being actively enforced following the Trump administration’s decision to dismantle the Biden-era AI Diffusion Framework earlier in 2026.
What Biden’s framework proposed — and why Trump scrapped it
Introduced in the final weeks of the Biden administration, the AI Diffusion Framework aimed to build a worldwide licensing system governing access to advanced AI processors, including strict export caps for countries outside America’s closest alliances. The proposal drew sharp criticism from the semiconductor industry. Nvidia — the world’s leading GPU manufacturer — argued the framework posed a threat to global innovation and cross-border research partnerships.
President Trump’s administration formally dissolved the framework in May 2026, describing it as carrying excessive regulatory burdens and unnecessarily straining US diplomatic relationships.
“The guidance reaffirms that NVIDIA’s sales and vetting process is correct — consistent with our existing approach, licenses are required to ship controlled products to PRC-headquartered companies.” — Nvidia spokesperson
The loophole critics say opened the door
Following the framework’s cancellation, critics alleged that Chinese companies had been exploiting an unaddressed gap — purchasing export-controlled chips through subsidiaries based in third countries where restrictions had not been explicitly applied. Former State Department technology policy official Chris McGuire stated publicly that Chinese firms had very likely been acquiring these chips at scale, and that the absence of updated BIS regulations meant such purchases were technically legal at the time.
The new guidance directly addresses this concern. BIS confirmed that shipments of restricted chips — including Nvidia’s top-tier Blackwell GPU lineup — to any company with Chinese headquarters or a Chinese parent entity are now prohibited, irrespective of where the subsidiary receiving the shipment is located. Companies that acquired chips through this window, however, are not required to cease using them.
Industry response
Nvidia confirmed it had already been operating in alignment with the newly clarified rules. Competitors AMD and Intel had not issued statements at the time of publication. Taiwan Semiconductor Manufacturing Company (TSMC), which fabricates chips for Nvidia and other clients, declined to comment. BIS also did not respond to media inquiries.
In a separate but related development, the Trump administration had previously permitted Nvidia to resume sales of its H200 chip to China in December 2025 — a significant relaxation of prior controls. The H200 is roughly six times more powerful than the H20, the most advanced chip previously cleared for Chinese export.
Conclusion
Washington’s latest guidance represents a targeted tightening of America’s semiconductor export control architecture — plugging a loophole that critics say allowed Chinese firms to quietly access restricted AI hardware through overseas channels.
While the move stops short of new regulatory rulemaking, it sends a clear signal that existing licensing requirements will be enforced globally, regardless of a company’s operational geography. As the US-China competition over AI dominance intensifies, the boundaries around advanced chip access are likely to remain a flashpoint.