Nvidia Resumes H200 Production for China, Accepting Orders Under New US Licensing Rules

Nvidia announced it is resuming production of its H200 GPUs for China, just 12 days after halting it, having secured purchase orders under new conditional licenses.

This is not a full-scale reopening but a 'conditional normalization' under strict U.S. export controls and limited approvals from China for select big tech firms.

Actual sales volume will likely be constrained by supply chain bottlenecks and policy risks, suggesting a more gradual, rather than explosive, impact on Nvidia's revenue.

Nvidia has officially announced it is restarting production of its H200 AI GPUs for the Chinese market, marking a significant turn in the ongoing U.S.-China tech saga.

This move comes just 12 days after reports on March 5th that Nvidia had halted China-bound H200 production and reallocated capacity to its next-generation 'Vera Rubin' chips. The quick reversal isn't a sign of indecision, but rather a response to a new set of conditions that have finally fallen into place, allowing sales to resume, albeit in a highly controlled manner.

The first major driver of this change is the shift in U.S. policy. What was once a near-blanket ban on high-end chip exports has morphed into a framework of 'strict conditional allowance.' This was formalized in mid-January when the U.S. Commerce Department established a case-by-case licensing system. The conditions are stringent: a reported 25% revenue share paid to the U.S. government, guarantees of sufficient supply for U.S. domestic needs, a ban on military use, and a sales cap at 50% of U.S. volumes. Nvidia's announcement confirms it has secured the necessary licenses under this new framework.

The second key factor is China's own 'managed opening.' After initially signaling a block on H200 imports in mid-January, Beijing shifted its stance. By the end of the month, it granted conditional import approvals to its tech giants—ByteDance, Alibaba, and Tencent—for an initial batch of over 400,000 units. This suggests China is not opening the floodgates but is selectively allowing strategic companies to acquire these critical chips, possibly while requiring them to also use domestically produced accelerators.

Finally, it's crucial to understand the supply-demand asymmetry. Reports from late 2025 indicated Chinese firms intended to order up to 2 million H200 chips, while Nvidia's available inventory was estimated at around 700,000. This gap is due to significant bottlenecks in the supply chain, particularly in advanced packaging (CoWoS) and high-bandwidth memory (HBM). Therefore, the production restart should be viewed not as a full recovery of demand, but as the 'monetization of some backlogged demand' under tight regulatory and supply constraints.

H200 GPU: A high-performance graphics processing unit designed by Nvidia for artificial intelligence workloads. The version for China is specifically configured to comply with U.S. export controls.
CoWoS (Chip-on-Wafer-on-Substrate): An advanced semiconductor packaging technology used to integrate multiple chips into a single package, crucial for high-performance GPUs but limited in production capacity.
Export Controls: Government regulations restricting the sale of specific technologies, such as advanced semiconductors, to foreign countries, often for national security reasons.

Nvidia Resumes H200 Production for China, Accepting Orders Under New US Licensing Rules

Nvidia announced it is resuming production of its H200 GPUs for China, just 12 days after halting it, having secured purchase orders under new conditional licenses.

This is not a full-scale reopening but a 'conditional normalization' under strict U.S. export controls and limited approvals from China for select big tech firms.

Actual sales volume will likely be constrained by supply chain bottlenecks and policy risks, suggesting a more gradual, rather than explosive, impact on Nvidia's revenue.

Nvidia has officially announced it is restarting production of its H200 AI GPUs for the Chinese market, marking a significant turn in the ongoing U.S.-China tech saga.

H200 GPU: A high-performance graphics processing unit designed by Nvidia for artificial intelligence workloads. The version for China is specifically configured to comply with U.S. export controls.
CoWoS (Chip-on-Wafer-on-Substrate): An advanced semiconductor packaging technology used to integrate multiple chips into a single package, crucial for high-performance GPUs but limited in production capacity.
Export Controls: Government regulations restricting the sale of specific technologies, such as advanced semiconductors, to foreign countries, often for national security reasons.