Nvidia expects to lose $5.5 billion after the U.S. government abruptly halted exports of its H20 AI chips to China and Hong Kong, the company disclosed this week. The new license requirement, effective as of mid-April 2025, is part of Washington’s broader push to restrict China’s access to advanced semiconductors over national security concerns.
The rule change follows a new interpretation of export controls and affects one of Nvidia’s top-selling chips. U.S. officials say the move aims to prevent Chinese use of powerful processors in supercomputing, raising the stakes in the ongoing tech cold war between the two nations.
AI Chip Sales to China Blocked
At the heart of the dispute is Nvidia’s H20 chip, an Artificial Intelligence processor designed to comply with earlier U.S. export restrictions. Despite those precautions, the Biden administration recently reclassified the chip, mandating a special export license for any sales to China, including Hong Kong. This abrupt policy shift has left Nvidia scrambling to cancel orders and adjust its inventory, triggering significant financial losses.
The U.S. government has justified the decision by citing national security concerns, specifically the potential for these advanced chips to be used in Chinese supercomputers. Nvidia confirmed that the license requirement "will be in effect for the indefinite future." The company added, "The [government] indicated that the license requirement addresses the risk that the covered products may be used in, or diverted to, a supercomputer in China."
The House Select Committee has now launched a formal investigation, claiming Nvidia’s H20 sales over the past year exploited a regulatory “loophole.”
Nvidia Defends Compliance and Economic Value
In a rare and swift response, Nvidia emphasised its strict adherence to U.S. export rules, saying, "The U.S. government instructs American businesses on what they can sell and where — we follow the government’s directions to the letter." The company also highlighted its role in supporting the U.S. economy, noting it pays substantial taxes, maintains a robust domestic workforce, and contributes to reducing the country’s trade deficit.
"NVIDIA protects and enhances national security by creating U.S. jobs and infrastructure, promoting U.S. technology leadership, bringing billions of dollars of tax revenue to the U.S. Treasury, and alleviating the massive U.S. trade deficit," the company stated.
Nvidia further clarified confusion around its second-largest billing market, Singapore. According to the company, while nearly $24 billion in sales were billed through Singapore last year, many of these orders originated from U.S. customer subsidiaries and were ultimately shipped to destinations like the U.S. and Taiwan—not China. "The associated products are shipped to other locations, including the United States and Taiwan, not to China," Nvidia said.
The Bigger Picture: Tech, Trade, and Geopolitics
This development comes amid growing concerns about China's rapid advancements in AI. In January, Chinese startup DeepSeek stunned the tech world with an AI model that rivalled U.S.-based offerings at a fraction of the cost—raising alarm bells in Washington.
Experts suggest the export restrictions are not just about individual companies like Nvidia, but part of a broader strategy to curb China's access to cutting-edge technology. However, some analysts predict the rules could shift again soon, as policymakers balance national security with the economic interests of the U.S. semiconductor industry.
Marc Einstein from Counterpoint Research called the $5.5 billion estimate reasonable and said, "While this is certainly a lot of money, this is something Nvidia can bear." He added, "I wouldn't be surprised to see some exemptions or changes made to tariff policy in the near future, given this not only impacts Nvidia but the entire U.S. semiconductor ecosystem."
Meanwhile, voices from the tech community, including industry insiders like Rui Ma, suggest a deeper decoupling may be inevitable. "It doesn't make any sense for any Chinese customer to be dependent on US chips," she warned, especially with an oversupply of data centers already in China.
