WASHINGTON: The United States has imposed a 25% tariff on a narrow range of advanced computing chips, including Nvidia’s H200 artificial intelligence processor, under a presidential proclamation that cites national security concerns tied to semiconductor supply chains. The measure, signed by President Donald Trump, applies to certain imports that are routed through the United States before shipment to other countries, including China, and took effect on Jan. 15.
The White House said the action was taken under Section 232 of the Trade Expansion Act of 1962 following a nine-month investigation into risks posed by reliance on foreign-made semiconductors and related equipment. The proclamation covers specified high-performance chips and related derivative products, and directs the Commerce Department to administer the policy and consider requests for additional exemptions within the framework set out in the order.
The tariff is limited by design, with the administration stating it does not apply to chips imported for use in U.S. data centers, startups, consumer devices that are not data center based, civil industrial applications, or public-sector uses. The White House also said the new tariff will not be cumulative with certain existing duties, and that the Commerce secretary has authority to grant further exemptions under the proclamation’s procedures.
The action arrives alongside a revised U.S. licensing approach that allows exports of Nvidia’s H200 chips to China under specified conditions. Under the updated requirements, exporters must certify that sufficient H200 supply remains available for U.S. customers, and Chinese purchasers must meet security-related conditions and restrictions on end use. The rules include third-party review requirements in the United States for chips intended for shipment under the permitted pathway.
New tariff linked to a tightly managed export channel
Nvidia’s H200, produced outside the United States and used in high-end AI training and inference workloads, has been at the center of shifting trade and technology restrictions. The administration’s approach pairs conditional export permission with a tariff applied to certain shipments that transit the United States, creating an added cost layer for some China-bound transactions while leaving large categories of domestic U.S. consumption outside the tariff’s scope.
Financial markets and technology companies reacted cautiously. Shares of major chipmakers moved modestly in the sessions surrounding the announcement, reflecting uncertainty over how narrowly the policy will be applied in practice and how frequently exemptions may be granted. AMD has said it complies with U.S. export controls, and the proclamation lists AMD’s MI325X among the chips subject to the tariff’s framework alongside the H200.
Chinese authorities have simultaneously signaled restrictions of their own. Chinese customs officials instructed agents to block the entry of Nvidia’s H200 chips, according to reporting citing people familiar with the guidance, and officials also told domestic technology companies not to purchase the chips unless absolutely necessary. The reported steps amount to a practical barrier to imports even in cases where U.S. licensing conditions would otherwise allow shipments.
China’s customs guidance adds friction to chip trade
In Washington, the sequence of decisions drew scrutiny from lawmakers who questioned the implications of allowing limited H200 sales to Chinese buyers while imposing a tariff tied to transshipment through the United States. Some members of Congress have raised concerns about advanced AI hardware reaching China, pressing for strict safeguards and enforcement. Administration officials have said the revised licensing channel includes conditions intended to control end use and requires certifications related to U.S. supply.
The tariff action reflects the breadth of presidential authority under Section 232, a tool that has been used in recent years to impose duties on imports on national security grounds. The administration has argued that U.S. dependence on overseas semiconductor manufacturing, including production concentrated in Taiwan, creates economic and security vulnerabilities. U.S. officials have cited estimates that the United States manufactures about 10% of the semiconductors it consumes, a figure used in the Section 232 rationale for the proclamation. – By Content Syndication Services.