Citing “significant economic and national security risk,” Donald Trump imposes a 25% tariff on Nvidia AI chips and other high-end semiconductors. The executive action, released Wednesday, specifically names the Nvidia H200 and AMD MI325X as targets. This follows a nine-month investigation concluding that the U.S. is too dependent on foreign manufacturers for its computational power.
The proclamation highlights a stark reality: the United States fully manufactures only roughly 10 percent of the chips it consumes. The administration views this imbalance as a vulnerability, particularly given the geopolitical tensions surrounding Taiwan, where most of these advanced chips are currently made. The tariffs are intended to serve as a strong incentive for companies to build foundries within U.S. borders.
Despite the aggressive posture, the White House has built in safety valves for the U.S. economy. The tariffs will not apply to chips imported for use by American datacenters, startups, or general consumers. This targeted approach attempts to shield the booming U.S. artificial intelligence sector from rising hardware costs while still penalizing the act of importing foreign-made goods.
The policy also acts as a de facto tax on China. Recent regulations force chips sold to China to be routed through the U.S. for testing by third-party labs. This detour exposes the shipments to the new 25% tariff. This aligns with Trump’s previous rhetoric regarding Beijing’s dominance in the chip sector and his broader strategy of using tariffs as leverage.
The Commerce Department, led by Howard Lutnick, retains the power to adjust these exemptions, adding a layer of uncertainty and flexibility to the rollout. While Nvidia has remained silent, AMD affirmed its compliance with export controls. This latest move sits alongside other recent tariffs on pharmaceuticals and vehicles, painting a picture of an administration eager to wall off the U.S. market to spur domestic growth.