The US Bureau of Industry and Security (BIS) has issued new guidance requiring export licences for advanced computing items destined for entities headquartered in mainland China or Macau — or whose parent companies are based there — even when those entities operate outside Chinese territory. The directive, published on May 31, 2026, has drawn a sharp rebuke from China's Ministry of Commerce, which accused Washington of abusing export controls and disrupting the global semiconductor supply chain.

What the BIS guidance actually says

The BIS clarification does not introduce entirely new restrictions; rather, it reinforces existing rules under the AI Diffusion rule framework, closing a perceived loophole that allowed Chinese-affiliated entities to procure advanced chips through overseas subsidiaries or third-party data centres. Licences are now explicitly required regardless of where a Chinese-headquartered or Chinese-parented company physically operates. Trade lawyers and industry insiders, however, said the practical fallout could be far more limited than the geopolitical tensions imply.

Why it matters for Chinese tech firms

Increasingly cut off from Nvidia's top-tier silicon at home, major Chinese technology companies — including the likes of Alibaba and ByteDance — have pivoted to data centres across Southeast Asia to secure the computing power needed to train next-generation AI models. The new guidance directly targets that workaround, potentially complicating offshore AI infrastructure strategies that have become central to these firms' competitive roadmaps. Firms advised by law practices such as King & Wood Mallesons are now reassessing their overseas procurement structures in light of the clarification.

The competitive backdrop

The guidance arrives amid sustained US efforts to prevent advanced AI hardware — particularly chips designed by Nvidia and manufactured by Taiwan Semiconductor Manufacturing Company (TSMC) — from reaching Chinese entities through indirect channels. Beijing has consistently framed such measures as economic coercion, while US officials maintain the controls are a national-security necessity. The tension has accelerated domestic chip development efforts inside China, though a meaningful capability gap with leading Western and Taiwanese foundries persists.

Market reaction and industry scrutiny

The announcement has triggered intense scrutiny across the semiconductor and cloud-computing industries, with legal teams and compliance officers at multinational firms reviewing exposure. Southeast Asian data centre operators that have quietly hosted Chinese AI workloads face particular uncertainty, as the guidance implies liability could extend to their upstream customers' corporate parentage. Analysts noted the rules place fresh pressure on regional cloud hubs in markets such as Singapore and Malaysia.

What's next

The degree to which BIS enforces the new guidance — and how aggressively it pursues licence violations involving offshore affiliates — will determine the real-world impact on Chinese AI development timelines. China's Ministry of Commerce is widely expected to respond with retaliatory trade measures or accelerated support for domestic chip champions. The next inflection point will be whether major Southeast Asian data centre operators begin turning away Chinese-affiliated clients to avoid US regulatory exposure.