Regulation

AI-generated text

U.S. Plans AI Chip Controls Targeting Remote Access Through Foreign Data Centers

The Trump administration is preparing export controls intended to prevent companies in China and other restricted countries from accessing advanced AI computing power through data centers abroad.

U.S. Plans AI Chip Controls Targeting Remote Access Through Foreign Data Centers

The Trump administration was preparing to introduce new export controls on artificial intelligence chips and data centers outside the United States on Monday. The measure is intended to prevent companies, particularly Chinese businesses, from circumventing U.S. export restrictions by remotely using AI infrastructure in other countries.

Under the interim rule, the Commerce Department would require licenses to export, reexport or transfer AI chips and servers to data centers anywhere in the world if those facilities may provide computing capacity to companies in China or other restricted countries.

The remote-access provisions target companies headquartered in countries in the U.S. D:5 group. The group includes China, Russia, Iran and North Korea, along with 18 other countries. Existing U.S. rules already require licenses to ship AI chips to companies based in those countries or owned by businesses there.

Capacity threshold and monitoring requirements

The new rule applies to data centers or networks capable of providing more than 12 million total processing performance, or TPP. The metric, used in U.S. export controls, corresponds roughly to the maximum computing power of 750 Nvidia H100 or H200 chips.

The restrictions also cover indirect access through intermediaries and brokers. Exporters and data center operators must monitor customers’ ownership structures and activities to ensure they are not controlled by parties from the D:5 group. Existing data centers providing services covered by the rule must obtain licenses within 90 days of its taking effect.

The Commerce Department would also introduce licensing requirements for exports, reexports or transfers of advanced semiconductors to Malaysia, Singapore, Indonesia, Thailand and Brazil. The department said those countries present a higher-than-normal risk that chips could be diverted to other destinations. The licensing requirement would not apply, however, to chips intended for U.S.-based companies. The rule also excludes commercial recommendation systems, such as those used in e-commerce and social media.

Extending existing controls

The interim rule describes the measure as an extension of existing controls, intended to prevent companies headquartered in the countries of concern from using advanced U.S. semiconductors to train advanced AI models. It also cites national security and foreign policy risks, arguing that such computing capacity could be used to train and run models supporting military or intelligence activities.

The move came days after, according to the source report, Nvidia co-founder Jensen Huang resigned amid a scandal involving an alleged Chinese spy linked to one of the chipmaker’s customers. The U.S. government has reportedly sought to place that customer, Megaspeed, on a special entity list. It believes the company bought thousands of Nvidia chips to provide computing capacity to Chinese businesses, thereby circumventing existing export restrictions. Reuters also reported the previous week that Chinese AI companies were rushing to secure computing capacity overseas amid fears of tighter U.S. controls.

The new rule also rescinds the Biden administration’s AI Diffusion Rule, introduced in January 2025. That framework sought to impose country-level limits on exports of advanced AI chips and restrict transfers of certain powerful AI models. In May of the same year, the Trump administration said it would discard the framework, arguing that it imposed unnecessary restrictions on U.S. companies and allies.