US Commerce Department moves to bring Nvidia chip remote access under export curbs

Control rather than location as the test

Export controls have traditionally depended on stopping chips at the border. Crypto Briefing reports that this approach falls short in the cloud era, since a firm can rent computing power in another country and run its workloads there without any hardware reaching China.

The proposed rules would therefore look at who directs and uses the equipment. Under the guidance described, China-headquartered firms would require licences for GPU chips, including Nvidia's Blackwell and H200 models, wherever those chips physically sit.

May guidance and a House bill

The plan rests on licensing guidance the department clarified in May 2026, which targeted gaps involving overseas subsidiaries and singled out Malaysia. A Chinese parent with a Malaysian subsidiary could, in theory, have reached hardware it was itself barred from purchasing.

Lawmakers have pursued a parallel course. In January 2026 the House approved the Remote Access Security Act by 369 votes to 22. The bill widens the definition of an export to cover cloud-based access to restricted hardware, so that using such a chip remotely could be treated as receiving it.

Scrutiny of Southeast Asian routes

Concern that restricted chips may reach Chinese companies through Southeast Asian intermediaries has added pressure for tighter rules. As of early November 2026, investigations were still running into approvals granted to Chinese firms tied to restricted entities, which places part of the bottleneck on the Chinese side of the approval process.

Open questions

• Will the Commerce Department formally adopt the proposed remote-access restrictions?
• How will regulators verify remote use of chips in overseas data centres?