Semiconductor major Nvidia is encountering unforeseen new U.S. export restrictions on its H20 chips.

Nvidia disclosed on Tuesday that the U.S. government has mandated an export license for its H20 AI chips destined for China, a requirement imposed indefinitely due to concerns the chips could be employed in Chinese supercomputers.

Nvidia expects related charges of $5.5 billion in its first quarter of fiscal year 2026, concluding on April 27, leading to an approximately 6% drop in the company's stock during extended trading hours.

The H20 represents Nvidia's most sophisticated AI chip permitted for export to China under the prevailing and preceding US regulations. Last week, NPR suggested that CEO Jensen Huang may have skillfully navigated potential new H20 limitations during a dinner held at President Donald Trump's Mar-a-Lago resort, partly by pledging Nvidia's investment in US-based AI data centers.

Intriguingly, Nvidia disclosed on Monday its intention to allocate hundreds of millions of dollars over the forthcoming four years towards the domestic fabrication of certain AI chips. Nevertheless, commentators were swift to highlight the company's pledge as lacking substantive specifics.

Several government officials had been asking for stricter rules on sending the H20 chip to other countries. This is because the chip was reportedly used to train AI models by the Chinese company DeepSeek. One of these models was the R1 "reasoning" model, which caused a lot of disruption in the U.S. AI market in January.

Nvidia refrained from issuing a statement.