The Trump administration is pushing to impose additional tariffs across the board on semiconductor chips and the information technology (IT) devices that use them, such as data center servers and laptops. The move is seen as an attempt to boost U.S. semiconductor manufacturing capacity. But there are concerns it could undercut itself by increasing the expense burden on domestic big tech companies that have staked everything on building artificial intelligence (AI) infrastructure.
Based on reports from major outlets including Politico and CNBC on the 27th, the Trump administration is reviewing a plan to levy comprehensive new tariffs on imported semiconductors and related finished goods. U.S. Commerce Secretary Howard Lutnick is reportedly considering a method of exempting tariffs in proportion to the scale of foreign corporations' investments in semiconductor manufacturing facilities in the United States. To reduce market shock, phasing in tariffs over a set period is also being seriously discussed. Citing multiple sources, Politico said internal talks are in full swing over the specific tariff rates and criteria for exemptions.
Building a semiconductor plant requires massive capital and years of time. If tariffs are imposed immediately, U.S. corporations that must rely entirely on Asian imported chips—such as from Taiwan or Korea, which account for more than 90% of cutting-edge chip output—would inevitably be hit. U.S. technology corporations, including big tech, also warned of this. Jonathan McHale, vice president for digital policy at the Computer and Communications Industry Association, whose members include major tech corporations such as Google and Meta, told Politico, "If you increase expense and reduce predictability, additional investment becomes difficult and it will also hinder the building of data centers (in the United States)."
There is also strong criticism that this conflicts with the Trump administration's core goal of securing U.S. leadership in AI. As America's giant technology corporations pour tens of trillions of won into building AI data centers, higher parts import expense could shrink the investment plans themselves. Trade experts noted, "Corporations that rely on overseas contract manufacturing, like Apple or Nvidia, could be put at a disadvantage in the global market due to rising costs."
Earlier, in January, the Trump administration imposed an additional tariff of as much as 25% on certain advanced AI semiconductor chips, including Nvidia's H200, based on Section 232 of the Trade Expansion Act. White House Spokesperson Kushi Desai said that day, "Reshoring semiconductor manufacturing facilities to the United States is President Trump's top priority," adding, "We have already secured investments of hundreds of billions of dollars that can be used for related policies." Michael Sobolik, a senior fellow at the Hudson Institute, told Politico, "Reducing supply chain risk is the core geopolitical task of our time," while analyzing, "Producing semiconductors at scale in the United States will entail massive expense."