Did Nvidia get permission to ship H200 chips to China?

A limited, uncertain reopening of exports

U.S. authorities granted Nvidia a narrow license to export a small number of its newest high‑end AI accelerators to customers in China. The company has publicly confirmed receiving the authorization, but the practical effect is still ambiguous: Washington’s permission is only one step in a two‑way trade that also depends on Chinese import approvals and on logistics that companies must still clear.

Even after the U.S. decision to allow restricted shipments, senior export‑control officials have said that, as of the latest reporting, no H200 systems had actually been sold or shipped into China. The lag reflects both the technical complexity of delivering the units under stringent compliance rules and ongoing geopolitical caution from purchasers and regulators on both sides.

Why the nuance matters

• Compliance friction: Export licenses carry strict conditions and monitoring requirements that slow transactions and make large deals impractical in the short term.
• Geopolitical signaling: The U.S. move is a calibrated loosening meant to balance alliance and national‑security concerns; Beijing’s agreement to accept shipments remains an independent, necessary step.
• Market implications: Nvidia’s data‑center business is a major driver of its record revenues, but constrained access to Chinese customers limits how quickly the company — and its global customers — can scale deployments in that market.

The situation illustrates how advanced semiconductors now sit at the intersection of trade policy, national security, and corporate strategy. Even with a license in hand, actual chip flows require both sides to follow through amid sustained political scrutiny.