The technology landscape shifted this week as Donald Trump imposes a 25% tariff on Nvidia AI chips, targeting the hardware at the center of the artificial intelligence boom. The order applies to specific high-performance chips, including the Nvidia H200 and AMD MI325X. It was issued under Section 232 of the Trade Expansion Act after an investigation found that foreign chip dependence threatens U.S. national security.
To prevent stalling the rapid progress of American AI, the administration included significant carve-outs. The tariff will not apply to chips imported for U.S. data centers, which consume the vast majority of these processors. Startups and government agencies are also exempt, shielding the domestic innovation ecosystem from immediate price increases.
The policy is particularly aggressive regarding China. Chips manufactured in Taiwan for the Chinese market must now be routed through the U.S. for testing. This detour allows the U.S. to apply the 25% tariff, effectively taxing the export of this technology to a strategic competitor. This move is part of a broader effort to maintain U.S. technological leadership.
The administration’s long-term goal is to reshore semiconductor manufacturing. With the U.S. currently producing only 10% of its chips, the White House views this as an unsustainable risk. The tariffs are meant to make foreign manufacturing less attractive and push companies to utilize incentives for building factories in the United States.
Market reaction saw slight declines in the stock prices of affected companies like Nvidia and AMD. However, the exemptions have seemingly prevented a broader sell-off. The administration has warned that this is an evolving situation, with the potential for broader tariffs if the industry does not take steps to increase domestic capacity.