Jensen Huang says Nvidia is in the process of restarting manufacturing of its H200 chips for shipments to China and it has received orders from “many customers”
Ina Fried /Axios:
Context & Ripple Effects
This follows Nvidia’s January statement that it had fired up H200 production pending resumed China sales and late-2025 reports that Chinese demand had already exceeded available output. The current update shifts the story from preparation to a stated manufacturing restart backed by customer orders.
China had also approved an initial batch of H200 imports for ByteDance, Alibaba and Tencent in January, while Nvidia had reportedly sought additional H200 capacity from TSMC. That sequence makes production availability—not merely customer interest—the immediate operational issue.
First-order effects
- Nvidia is restarting H200 manufacturing for China-bound shipments, turning previously anticipated sales into production work and order fulfillment.
- Chinese H200 customers that have placed orders gain a clearer supply path, though shipment timing and volumes are not specified.
Second-order effects
- The reported order base strengthens Nvidia’s case for allocating more H200 capacity to China; prior reporting that demand outstripped output means that allocation can compete with other manufacturing commitments.
- Foundry and component planning may need to adjust if the restart becomes sustained: Nvidia had previously approached TSMC about ramping H200 output in response to Chinese orders.
Third-order effects
- If China-bound H200 production continues, export-policy clearance and semiconductor capacity planning will become more tightly coupled: demand can return faster than the supply chain can reallocate output.
- The episode fits a more segmented AI-compute market, in which vendors tailor product availability and capacity decisions to the permissions and demand conditions of individual regions.
The trend: AI-chip supply is increasingly being planned around regional market access, with policy changes translating directly into capacity allocation decisions.