A detailed look at the new executive order reveals how Donald Trump imposes a 25% tariff on Nvidia AI chips. The policy targets specific high-performance semiconductors, such as the Nvidia H200 and AMD MI325X, under the authority of Section 232 of the Trade Expansion Act.
The order is complex, featuring significant exemptions. Chips imported for U.S. data centers, startups, and government use are not subject to the tariff. This strategic decision aims to shield the U.S. tech industry from the costs of the trade war, ensuring that American AI development continues apace.
The tariff’s main mechanism involves the China trade. Chips manufactured in Taiwan and sold to China must now pass through the U.S. for testing. This mandatory stopover classifies the goods as imports, triggering the 25% duty. It is a way to impose costs on a strategic rival without violating international export rules.
The administration’s motivation is clear: reduce reliance on foreign chips. With the U.S. producing only 10% of its own semiconductors, the White House views the supply chain as a security risk. The tariffs are designed to encourage companies to move production to the U.S., thereby securing the nation’s technological infrastructure.
Investors reacted cautiously, with tech stocks trading slightly lower. The administration has indicated that the policy is adaptable, with the Commerce Secretary holding the authority to modify exemptions. The threat of broader tariffs remains a key part of the administration’s strategy to force a change in industry behavior.