US Targets AI Chip Cloud Loophole With New Data Center Rules

US AI chip restrictions now target remote computing access through overseas data centers, with new licensing rules for five countries.

Nov 17, 2026
US Targets AI Chip Cloud Loophole With New Data Center Rules

New US export controls target remote access to advanced AI computing power hosted in overseas data centers. Image: Nvidia

We may earn from vendors via affiliate links or sponsorships. This might affect product placement on our site, but not the content of our reviews. See our Terms of Use for details.

An advanced AI chip can stay in a data center overseas while a customer thousands of miles away rents its computing power. Washington is moving to tighten oversight of who gets that access.

The US Commerce Department announced new restrictions on November 16 targeting remote access to advanced American AI computing systems. The rule also adds licensing requirements for certain semiconductor shipments to Singapore, Malaysia, Indonesia, Thailand, and Brazil.

The measures could complicate cloud operations and chip procurement for companies serving international AI customers. Existing export controls already restrict physical shipments; the new rules also target certain arrangements involving remotely supplied computing power.

How the new AI chip controls work

The Commerce Department’s Bureau of Industry and Security (BIS) issued an interim final rule titled Addressing Remote Access to Advanced Semiconductors. It introduces licensing controls on certain exports, reexports, and transfers of advanced computing equipment.

Advertisement

Under the new end-use restrictions, covered equipment requires a license when exporters know it will be used in data centers providing more than 12 million aggregate total processing performance (TPP) to specified restricted-country companies. TPP measures computing capability under US export regulations. The threshold can include capacity shared across linked data centers.

The controls cover access supplied directly or indirectly to companies headquartered in Macau or US Country Group D:5 destinations, including China and Russia. Companies with ultimate parents headquartered in those locations are also covered.

Separately, certain advanced-chip shipments to the five named countries face additional licensing because of diversion risks identified by BIS. Earlier investigations have raised concerns about chips being routed through third countries, including an alleged $300 million AI server smuggling scheme involving shipments through Malaysia and Singapore.

The restrictions are not a blanket ban. Qualifying US-headquartered companies are exempt from the additional five-country licensing requirement if reporting conditions are met. Certain systems used exclusively for commercial recommendations are also excluded from the remote-access control, subject to specific conditions, including reporting requirements. Generative AI systems do not qualify for that exclusion.

Advertisement

Overseas data centers face closer scrutiny

Semiconductor exporters and data center operators face additional customer-screening responsibilities, including checks on ownership and control. The federal rule follows a separate reported Nvidia GPU monitoring initiative involving prospective cloud partners in Southeast Asia.

Enterprise IT teams may need to examine GPU suppliers, cloud contracts, and hosting arrangements for licensing exposure. Changes in how equipment is used can also require authorization, even when the hardware remains at the same facility.

The new requirements arrive as Southeast Asian data center investment expands. In a November 16 statement, Singapore’s government said it was consulting affected companies and engaging US authorities about the measures’ scope and implementation.

The rule is scheduled to take effect upon Federal Register publication on November 19, 2026. BIS will accept public comments for 60 days after publication. Certain existing covered data center arrangements have 90 days from the effective date to obtain required licenses.

Advertisement

For companies relying on overseas GPUs, compliance now involves more than where chips are shipped. Who ultimately uses their computing power also matters, while BIS licensing decisions will help determine the effects on international cloud operations.

Want to learn more AI tips, tricks, and prompting techniques? Let us teach you How to Talk to AI for free! Try our six-minute course at The Neuron Academy, our practical learning platform designed to help professionals use AI more confidently at work.

Learn a few simple ways to write better prompts and get more useful results from AI, or browse our other AI course for free for seven days. Check out all the lessons here →