Nvidia H200 Exports Allowed by US, China Restricts Immediate Imports

In Focus

• US allows Nvidia H200 chip exports to China under strict export rules
• Chinese authorities instruct firms not to import H200 chips unless necessary
• Global AI supply chains face reconfiguration amid geopolitical tensions

The United States has eased restrictions on Nvidia H200 exports, allowing the advanced artificial intelligence chips to be sold to China under a tightly controlled regulatory framework. The policy marks a departure from previous blanket bans on high-end AI hardware exports, reflecting a more measured approach to balancing commercial opportunities with national security concerns.

While the Trump administration has approved shipments under strict licensing and verification conditions, Chinese authorities have instructed domestic technology companies not to import the chips unless necessary. As per a report by Reuters, this will create what multiple sources describe as a temporary ban, although the situation could progress in the coming months.

Revised AI Chip Export Rules and Licensing Framework

Under the updated export rules, Nvidia can ship H200 processors to Chinese customers only after meeting several compliance requirements. These include third-party verification of the chip’s performance, confirmation that sufficient domestic supply exists in the US, and assurances that the chips will not be used for military purposes.

The H200 export conditions also impose volume limits tied to Nvidia’s US sales and require buyers to demonstrate robust security controls. Each shipment remains subject to case-by-case licensing, reinforcing oversight rather than unrestricted access.

The regulatory adjustment reflects a shift from outright prohibition to conditional engagement. US officials have indicated that the H200, while powerful, is not Nvidia’s most advanced architecture, enabling policymakers to maintain a technological lead while permitting commercial transactions.

Supporters argue that this approach allows US companies to generate revenue and sustain research investment without transferring their advanced capabilities. Critics, however, caution that enforcement may prove challenging once chips enter commercial data centers, given their potential for dual-use applications.

China’s Restrictions and Market Implications

Chinese technology companies have reportedly been instructed to restrict purchases of H200 chips, with limited exemptions for academic and research partnerships. It remains unclear whether the restrictions apply to existing orders or only to new purchases.

Analysts suggest that Beijing’s motives may include supporting domestic AI chip development, testing restrictions, or leveraging the H200 in negotiations with Washington ahead of President Trump’s April visit to Beijing.

Demand for the H200 chips remains exceptionally high. Chinese companies reportedly placed orders for more than two million chips, far exceeding Nvidia’s inventory of 700,000 units. The H200 delivers approximately six times the performance of the older H20 chip, making it a highly desirable option for training large-scale AI models.

Re-entry into the Chinese market could generate significant revenue for Nvidia and the U.S. government, which would collect a 25% export fee, but geopolitical sensitivities and shipment restrictions limit the potential upside.

Experts also note that China may be using the restrictions as a bargaining tactic. Beijing appears aware that the United States is eager to sell AI chips and may try to extract concessions in exchange for license approvals, a strategy that could influence upcoming trade and technology discussions between the two nations.

Geopolitical Impact on the Semiconductor Industry

The approval of Nvidia H200 chip exports to China and the subsequent restrictions reflect the complex intersection of technology, geopolitics, and trade policy. For Nvidia, conditional access to China offers revenue potential, although caps, compliance requirements, and enforcement challenges limit immediate gains.

Global AI supply chains are increasingly segmented, with the United States and allied countries accessing the latest-generation chips while China receives constrained access under strict oversight.

Investors and enterprise technology providers may view the policy as reducing regulatory uncertainty, but the ongoing geopolitical tensions and enforcement risks ensure that strategic planning in the AI hardware sector remains highly sensitive.