Following prolonged speculation, the Trump administration has officially introduced a tariff targeting specific semiconductors. This new levy primarily affects certain advanced AI chips, notably the Nvidia H200, which are slated for export to China.
On Wednesday, President Donald Trump formalized a 25% tariff through a signed proclamation. This tariff applies to advanced AI semiconductors manufactured abroad that transit through the U.S. en route to international customers.
This development officially enacts a crucial aspect of the U.S. Department of Commerce’s prior decision in December, which authorized Nvidia to commence shipping its H200 advanced AI chips to pre-approved clients in China. The tariff also encompasses offerings from other manufacturers, such as the AMD MI325X.
Despite the imposition of these tariffs, Nvidia publicly welcomed the policy, as it enables the company to supply the chips to authorized customers.
An Nvidia spokesperson conveyed to TechCrunch, “We commend President Trump’s initiative to empower America’s chip industry, fostering competition and bolstering high-paying jobs and domestic manufacturing. The provision of H200 chips to commercial clients, sanctioned by the Department of Commerce, represents a judicious compromise highly beneficial to America.”
Significant demand exists for these H200 semiconductors, with Nvidia reportedly contemplating increased production to address a surge in initial orders from Chinese firms.
However, demand is only part of the equation; the regulatory stance adopted by the Chinese government toward these imports presents another critical variable.
China confronts a predicament both analogous to and distinct from that of the U.S. concerning semiconductor manufacturing and the worldwide AI competition. While Beijing aims to bolster its indigenous semiconductor sector, it simultaneously seeks to avoid lagging behind global competitors during its domestic technological development phase.
As reported by Nikkei Asia, China’s central government is actively formulating regulations to govern the volume of semiconductors Chinese companies can acquire from foreign sources. Such a policy would permit some acquisition of Nvidia’s chips, marking a significant departure from the nation’s prior resistance to these chip imports.
The executive order issued on Wednesday explicitly exempts chips imported into the U.S. for domestic use in research, defense, or commercial applications.
The proclamation underscored, “The United States presently produces only around 10% of its necessary chips, leading to substantial reliance on international supply chains. This reliance poses a considerable economic and national security vulnerability.”