A senior U.S. trade official has confirmed that only a minimal volume of Nvidia’s H200 artificial intelligence chips has been exported to China and Hong Kong. During a recent congressional hearing, Under Secretary of Commerce for Industry and Security Jeffery Kessler stated that while some shipments have occurred under specific licenses, the total quantity remains very small.

This development marks a shift in the trade landscape for Nvidia, which had previously excluded potential Chinese revenue from its financial forecasts. Despite the resumption of limited shipments, the company continues to navigate a complex environment defined by ongoing trade tensions and strict U.S. export controls. The H200, part of the older Hopper generation, remains subject to rigorous case-by-case review by the U.S. government to ensure compliance with national security standards.

While the U.S. has begun issuing licenses for these chips, the long-term outlook for large-scale exports remains uncertain. Applicants for these licenses must undergo thorough inspections and meet stringent security requirements, with many applications still facing outright denial. For Chinese firms, the ability to access these chips is critical, as domestic alternatives currently struggle to match the performance levels required for advanced AI training.

Key Takeaways

• U.S. officials confirmed that only a small quantity of Nvidia H200 chips has been shipped to China and Hong Kong.
• Export licenses for the chips are granted on a case-by-case basis, subject to strict national security inspections.
• Despite the limited resumption of sales, many license applications continue to be denied by the U.S. government.

Editor’s Analysis & Impact

The limited resumption of Nvidia H200 shipments to China represents a delicate balancing act between U.S. national security interests and the commercial realities of the global semiconductor market. By allowing a trickle of older-generation chips into the Chinese market, the U.S. maintains a degree of leverage while preventing a total technological decoupling. However, the reliance on case-by-case licensing suggests that the U.S. government intends to keep a tight leash on China’s AI development capabilities. For Nvidia, while any revenue from the Chinese market is a positive development, the company remains largely dependent on its newer Blackwell architecture for growth in other regions. The broader implication is that the ‘chip war’ is evolving into a highly granular, bureaucratic process where access is used as a strategic tool rather than a blanket prohibition.

Frequently Asked Questions

Q: Why are Nvidia H200 chips subject to U.S. export restrictions?
A: The U.S. government restricts the export of high-performance AI chips to China due to concerns that these technologies could be utilized for military purposes.

Q: Are Chinese companies able to produce their own AI chips?
A: Yes, Chinese firms are developing domestic alternatives; however, these products are currently considered inferior to Nvidia's technology for advanced AI training tasks.