Nvidia announced on Tuesday that it would take $5.5 billion in charges. This is after the U.S. government limited exports of its H20 artificial intelligence chip to China. China is a key market for one of its most popular chips.
U.S. export controls have focused on Nvidia’s AI chips. U.S. officials have moved to prevent the sale of the most advanced chips to China. The U.S. aims to maintain its lead in the AI race. After these controls were put in place, Nvidia started designing chips that would come as close as possible to U.S. limits.
Nvidia shares fell about 6% in after-hours trading.
The H20 is currently Nvidia’s most advanced chip for sale in China. It is central to Nvidia’s efforts to remain engaged with China’s booming AI industry. Reuters reported in February that Chinese companies like Tencent, Alibaba, and ByteDance had been increasing orders for H20 chips. This was due to booming demand for low-cost AI models from startup DeepSeek.
The H20 chip is not as fast at training AI models as Nvidia’s chips for sale outside China. However, it is competitive with some of those chips at inference. Inference is where AI models provide answers to users. Inference is quickly becoming the biggest part of the AI chip market. Nvidia CEO Jensen Huang stated last month that Nvidia is well positioned to dominate that shift.
On Tuesday, Nvidia said that the U.S. government is restricting H20 sales to China. This is because of the risk that the chips could be used in a supercomputer. The H20 has lower computing capabilities than other Nvidia chips, but its ability to connect to memory chips and other computing chips at high speeds remains high.
Those memory and connectivity aspects could make the H20 useful in building supercomputers in China. The U.S. has restricted the sale of chips for use in supercomputers in China since 2022. The Institute for Progress, a nonpartisan think tank, argued on Tuesday for restricting the H20 chips. They wrote that Chinese firms were likely already building such systems.
The group wrote, “At least one of the buyers, Tencent, has already installed H20s in a facility used to train a large model, very likely in breach of existing controls restricting the usage of chips in supercomputers exceeding certain thresholds. DeepSeek’s supercomputer used to train their V3 model is also likely in breach of the same restrictions.”
On Thursday, a Tencent spokesperson said the Institute for Progress’ conclusions were inaccurate.
Tencent said in a statement, “We have not violated any laws and have not built any ‘supercomputers’. Any claim to the contrary is categorically false.”
Responding to Tencent’s comments, Tim Fist, one of the authors of the institute’s report, said the institute based its conclusions on widespread AI industry practices around installing Nvidia servers. Fist said it is possible Tencent complied with U.S. rules. However, doing so would involve accepting performance limitations and extra installation costs.
Nvidia said on Tuesday that the U.S. government informed it on April 9 that the H20 chip would require a license to be exported to China. On April 14, the U.S. government told Nvidia those rules would be in place indefinitely.
It is unclear how many, if any, of those licenses the U.S. government might grant.
Nvidia declined to comment beyond its filing. The U.S. Department of Commerce did not immediately return a request for comment. The department oversees U.S. export controls.
The $5.5 billion in charges are associated with H20 products for inventory, purchase commitments, and related reserves, Nvidia said. Nvidia announced on Monday its plan to build AI servers worth as much as $500 billion in the U.S. over the next four years. They will do this with help from partners such as TSMC. This is in step with the Trump administration’s push for local manufacturing.
Click here for more Technology news.