Nvidia, a preeminent semiconductor titan, finds itself ensnared in an unanticipated entanglement with novel US export strictures concerning its H20 chip.

In a Tuesday regulatory filing, Nvidia disclosed notification from the U.S. government stipulating the perpetual requirement of a license for the export of its H20 AI chips to China, citing the perceived "risk that the [H20] may be used in […] a supercomputer in China."

Nvidia is anticipating $5.5 billion in related expenditures for the first fiscal quarter of 2026, concluding on April 27, 2026, precipitating an approximately 6% decline in its stock price during extended trading.

The H20 represents Nvidia's most advanced AI chip currently permissible for export to China under prevailing and historical export controls.

Last week, NPR reported that CEO Jensen Huang may have successfully negotiated an exemption for new H20 restrictions during a dinner at Donald Trump's Mar-a-Lago resort, a potential quid pro quo purportedly involving a pledge by Nvidia to invest in AI data centers within the United States.

In confluence with this trajectory, Nvidia declared on Monday its intention to commit hundreds of millions of dollars over the next four years towards the domestic fabrication of a subset of its AI chips, a pronouncement that critics promptly flagged as lacking specific granular detail.

Senior government officials had called for more stringent export controls, citing suspicions that the H20 chip had been employed in the training of models by the Chinese AI startup DeepSeek, including the R1 "reasoning" model that sent ripples through the US AI market in January.

Nvidiaは、当該事案に関する一切の言及を差し控えるという断固たる姿勢を貫いた。