U.S. considers semiconductor tariffs on servers and electronics

Traders Union
All News
Economic Data
U.S. semiconductor tariffs

Highlights

New chip tariffs could raise AI infrastructure costs.
Data centers remain dependent on Asian chip supplies.
U.S. manufacturing capacity will take years to expand.

According to POLITICO, officials are weighing duties that could extend beyond imported chips to products containing them, including data-center servers, laptops, and gaming consoles. One approach would link tariff relief to commitments by foreign semiconductor companies to expand manufacturing in the United States, while officials are also considering a phase-in period. The framework remains under discussion and could change before any final decision.

Data centers face a new cost risk

The timing is particularly sensitive for U.S. technology companies. Demand for advanced semiconductors has surged as Amazon, Google, Meta, and other companies commit hundreds of billions of dollars to AI infrastructure and large data-center campuses. Industry representatives are lobbying for exemptions that would protect data centers and other domestic uses from the proposed duties.

The central problem is supply. The United States cannot currently manufacture enough leading-edge chips to satisfy domestic demand. Taiwan alone accounts for more than 90% of the world's most advanced semiconductor production, while South Korea and Malaysia are also important links in U.S. technology supply chains. Building comparable capacity domestically requires billions of dollars and years of construction.

Commerce Secretary Howard Lutnick favors a system under which companies could import a certain volume of chips without duties based on how much manufacturing they commit to the United States. Officials have not settled on tariff rates, and country-specific quotas and rates are among the options being discussed.

AI expansion meets industrial policy

For companies building AI infrastructure, broader tariffs could increase the cost of servers before domestic semiconductor capacity is ready to replace imports. Nvidia and AMD could also face additional pressure because they design chips in the United States but rely heavily on overseas manufacturers for production.

The gap cannot be closed quickly. TSMC has committed $265 billion to manufacturing facilities in Arizona, yet the company expects the U.S. plants to represent only about 30% of its most advanced capacity after the planned expansion is completed. Much of that capacity remains years away.

That creates a difficult transition period. Tariffs could encourage additional semiconductor investment over the long term, but in the near term they would increase the price of components that U.S. companies cannot readily source domestically. The result could be higher data-center construction costs or delays to planned projects.

The stakes for the U.S. AI buildout

The policy debate reaches beyond semiconductor trade. Data centers require enormous quantities of GPUs, memory, and server equipment, making chip availability central to the U.S. effort to expand AI computing capacity.

Washington therefore faces a sequencing problem: domestic semiconductor production is expanding, but not fast enough to replace imports immediately. If tariffs arrive before new factories reach scale, the cost of accelerating U.S. chip manufacturing could fall directly on the AI infrastructure the administration also wants to strengthen.

We also reported Trump expanded tariffs across 60 economies.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.