AI policy
Washington closes the subsidiary loophole — and the chip controls just got harder to route around
Guidance tying the rules to a company's parent, not its address, redraws where advanced AI chips can legally go.
On 1 June 2026 US guidance affirmed AI-chip licensing applies to Chinese firms' overseas units too.
On 1 June 2026, the US Commerce Department issued guidance affirming that licensing requirements for advanced AI chips apply to any business headquartered in or parented by a Chinese company — not just operations physically inside China. It is a clarification with teeth: it closes what had been a usable workaround.
The loophole that just closed
Until now, a Chinese company's overseas subsidiary — registered in a third country — could plausibly sit outside the controls that targeted China by geography. The new guidance reframes the test around ownership: if a Chinese firm is your parent, the rules follow you, wherever you are incorporated. That is a meaningful change in how the export-control regime is scoped.
Who it hits
Most directly, Nvidia — including its Blackwell-class chips — now faces licensing friction on sales to the foreign units of Chinese firms, a customer set that had been a grey area. The broader signal is to every chipmaker and cloud provider weighing sales to entities with Chinese parentage: the compliance question is no longer 'where is the buyer' but 'who ultimately owns them'. Expect that test to shape commercial decisions well beyond this one guidance note.
Sources
- US says ban on AI chip shipments applies to Chinese firms outside China — Al Jazeera, 1 June 2026