Wednesday, November 18, 2026
Tech

U.S. Moves to Restrict Remote Access to Advanced AI Chips Through Overseas Data Centers

A Commerce Department interim final rule would require licenses for certain overseas data centers serving companies in restricted destinations and add controls for five other countries.

Close-up of a semiconductor wafer showing rows of microchip dies
CC0 illustrative photograph of a semiconductor wafer; it does not depict a specific chip or facility covered by the U.S. rule.

WASHINGTON, November 16, 2026 — The Commerce Department’s Bureau of Industry and Security has issued an interim final rule designed to prevent companies in restricted destinations from accessing advanced U.S. computing chips through data centers located in third countries.

The measure addresses a gap that U.S. officials say has emerged as cloud computing has made it possible to rent computing capacity remotely instead of importing physical chips. The government already restricts exports of some of the most advanced semiconductors to China and other destinations, but overseas data centers can give customers access to processors without the hardware being shipped directly to them.

BIS issued the interim final rule on Sunday, Nov. 15, and the agency and news organizations discussed it publicly on Monday, Nov. 16. Its accompanying regulatory text says it will take effect on the date it is published in the Federal Register. The public-inspection document available at the time of this report still contained placeholders for that date and the comment deadline, so the rule was not yet in force. The document provides for a 60-day public-comment period after publication.

How the new controls would work

The rule would require a license for exports, reexports or transfers of covered advanced computing items when the exporter knows the equipment will be used in a data center that provides significant computing power to a company headquartered in Macau or a destination listed in the Commerce Department’s Country Group D:5. The control also applies when a company’s ultimate parent is headquartered in one of those destinations.

The rule defines the threshold as more than 12 million aggregate total processing performance points, a measure of computing capacity that can apply to one data center or to linked facilities. It also covers indirect access, including arrangements in which one company provides computing services for another company that is ultimately serving a restricted customer.

Data centers already providing covered services before the rule’s publication would have 90 days from the effective date to obtain a license if they want to continue. The document requires data centers, foreign chip manufacturers and exporters to conduct due diligence on customers, including examining direct and indirect ownership or control.

Five countries face additional licensing requirements

The rule would also require licenses for advanced semiconductor exports, reexports or transfers to Indonesia, Malaysia, Singapore, Thailand and Brazil, which the department identifies as destinations presenting a higher-than-normal risk of diversion. The requirement would not apply to certain shipments destined for companies headquartered in the United States, or whose ultimate parent is headquartered there, if specified reporting requirements are met.

Commercial recommendation systems, such as tools used by online retailers and social media companies to rank or suggest products and content, are excluded from the remote-access control. The rule also formally rescinds the Biden administration’s AI Diffusion Rule, which the Commerce Department says was never enforced.

Why Washington is acting now

The new measure follows concerns that companies could obtain access to U.S.-designed processors through data centers in Southeast Asia and other locations. Reuters and the Associated Press reported that U.S. officials see remote computing as a potential route around physical export restrictions. Allegations involving particular companies and possible diversion remain allegations unless established by evidence or official findings.

The administration says the controls are intended to protect technology that could be used in advanced military, intelligence or cyber capabilities. Technology companies and data-center operators, meanwhile, will need to assess how the rules affect customer screening, contracts, compliance systems and access to computing services.

What it means for U.S. readers

The rule could affect U.S. chipmakers, cloud providers, data-center operators and international customers that rely on American processors. It may also influence the cost and availability of advanced computing capacity in countries that serve as regional technology hubs. The measure is part of a wider effort to align control of physical exports with control of remote access, but its final impact will depend on the published effective date, licensing decisions and the government’s response to public comments.

Sources and reporting credit

The Fact Bridge Editorial Team covers U.S. politics, business, technology, India and world affairs. Our reporting is grounded in official documents, direct statements and clearly attributed reporting from credible sources. We distinguish verified facts from allegations and analysis, and correct substantive errors in line with our Corrections Policy.

Leave a Reply

Your email address will not be published. Required fields are marked *