The US Commerce Department intends to prevent China from remotely controlling Nvidia chips located beyond its borders.
For years, export controls have rested on a straightforward premise: keep the chip from crossing the border. But with cloud computing, the chip does not need to move.
What the Commerce Department is targeting
Advanced AI chips can be used in much the same way. A company can rent computing capacity at a data center in another country and run workloads there from anywhere, without any hardware ever reaching Chinese soil.
Washington now wants to close that gap. The plan would apply restrictions not just to physical shipments but also to remote control of Nvidia hardware located abroad.
The move builds on licensing guidance the department clarified in May 2026. Under that guidance, companies headquartered in China must obtain licenses for GPU chips, including Nvidia’s Blackwell and H200 models, even if the chips are outside China.
The guidance was intended to address loopholes involving overseas subsidiaries, with Malaysia specifically highlighted. In theory, a Chinese parent company with a Malaysian subsidiary could have accessed hardware it was prohibited from purchasing directly.
In practical terms, the rules are shifting attention from a chip’s location to who is using it.
How the rules got here
The push to restrict remote access did not emerge without warning. In January 2026, the House passed the Remote Access Security Act by a 369-22 vote.
The bill broadens the definition of an export to include cloud-based access to restricted hardware. In other words, remotely logging into a restricted chip could be treated the same as having it shipped to you.
Scrutiny has also increased over how restricted chips may be reaching Chinese companies through Southeast Asian intermediaries. Investigations into possible diversion of chips through the region have added pressure to tighten the rules.
As of early November 2026, investigations were underway into approvals granted to Chinese companies linked to restricted entities.
Congress is also pursuing further measures, including chip-tracking requirements, as part of the FY2027 defense authorization bill.
The Hong Kong wrinkle
About 600,000 Nvidia H200 chips were reportedly stockpiled in Hong Kong as of November 2026.
The chips were reportedly awaiting import approval from mainland China amid ongoing geopolitical tensions. In that case, the bottleneck lies with the Chinese approval process, not the US one.
What this means for Nvidia and the chip trade
For Nvidia, the overall direction is clear, even though the final details remain uncertain. Each new rule reduces the number of customers able to use its most advanced products without government approval.
Cloud providers face a particular risk. If remote access is considered an export, data center operators outside China may need to vet customers much more carefully, including tracing corporate ownership through subsidiaries.
Several developments warrant attention. First, will the Remote Access Security Act pass the remaining stages of the legislative process and become binding law?
Second, which chip-tracking provisions will remain in the FY2027 defense authorization bill?
Third, what will ongoing investigations into approvals for firms connected to restricted entities find?
Then there are the roughly 600,000 H200 chips in Hong Kong.
The broader message for the industry is that the definition of an export is widening. A chip’s physical location once told the whole story. Regulators are increasingly focused on control and ownership.
Source: cryptobriefing.com. Adapted by NBTC.
Add A Comment
Comments are closed.