Nvidia Secures US License to Export Limited H200 Chips to China, But Challenges Remain

American semiconductor giant Nvidia recently received a rare export license from the United States government allowing it to ship a small number of its H200 artificial intelligence chips to customers in China, marking a cautious step in the company’s efforts to re-enter the world’s largest semiconductor market. However, significant hurdles remain before meaningful volumes of AI processors flow into China, even as geopolitical and trade tensions continue to shape the global technology landscape.

The US Department of Commerce granted the license under specific conditions: all chips must undergo inspection in the United States before shipment, and a 25 % duty applies to the exports. Despite this regulatory clearance, Nvidia has yet to record any actual sales of the H200 to Chinese buyers, and it has excluded China revenue from its latest quarterly outlook.

A Modest Step Forward

The H200 — Nvidia’s second-most powerful AI accelerator behind its flagship Blackwell series — was previously restricted from export to China due to longstanding US export controls designed to limit Beijing’s access to advanced computing technology. That changed late last year, when the US administration shifted strategy and began issuing licenses on a case-by-case basis.

Under the new rules, certain high-performance AI chips can be exported if companies meet stringent requirements: domestic supply must be sufficient to satisfy US customers first, independent third-party testing must verify chip capabilities, and Chinese purchasers must demonstrate compliance measures aimed at preventing military or prohibited uses. There are also implicit caps on quantities relative to domestic shipments.

Though the license represents a policy shift, none of the H200 chips subject to this program have yet entered China. According to a senior US export enforcement official, zero H200 units have been sold to Chinese customers as of late February 2026, even after the initial licensing decision was made. The official noted that enforcement efforts to prevent smuggling — and to ensure compliance with export restrictions — remain vigorous.

Geopolitics and Trade Tensions

Nvidia’s limited export success comes against the backdrop of heightened tensions between the United States and China over semiconductor technology. Advanced chips are seen as strategic assets in the race for AI dominance, and policymakers in Washington have long worried that unfettered access to cutting-edge processors could bolster China’s military and surveillance capabilities.

The licensing move reflects an attempt to balance national security concerns with economic opportunities. China is a massive potential market for AI hardware: Nvidia’s CEO has previously estimated its total addressable market there could reach roughly $50 billion over time. However, Beijing has been cautious in granting formal approval for imports, even after the US eased some restrictions.

China’s domestic semiconductor industry — including firms such as Huawei Technologies, Cambricon Technologies, and other emerging AI chip makers — has been aggressively expanding, partly in response to export barriers. These companies are receiving significant domestic support, complicating Nvidia’s efforts to regain market share.

Why No Sales Yet?

Industry analysts and export control experts point to multiple factors explaining why the license hasn’t yet translated into actual shipments:

• Policy uncertainty in both countries: Even with a license, final approval of specific transactions remains complex and protracted, requiring coordination with Chinese regulators and compliance verification.

• US safeguard mechanisms: Chips must be inspected in the US before export, adding logistical steps and potential delays.

• Chinese import controls: Beijing hasn’t fully signed off on the acceptance of these chips into its market, possibly due to reciprocal political signaling or strategic industrial policies.

• Tariffs and cost structures: The 25 % duty could dampen commercial enthusiasm, as importing these high-end chips becomes more expensive for buyers.

  
  

All of these factors contribute to a situation where permission does not yet equal execution.

Broader Industry Implications

The situation highlights the complex interplay between technology leadership, national security, and global commerce in the AI era. Nvidia’s dominant position in AI accelerators means that restrictions on its ability to sell in key markets have ripple effects across the tech ecosystem, influencing everything from data-center buildouts to AI research priorities.

For Nvidia, limited access to China — historically one of its largest markets — risks ceding ground to competitors and local players. According to filings, the inability to sell advanced chips to China previously led to substantial inventory charges and lost sales, with competitors expanding their ecosystems in the interim.

On the policy side, the licensing framework represents an evolving approach: instead of outright bans, regulators are experimenting with controlled access to balance security with economic interests. How this balance holds up will be crucial for other technology firms navigating US-China relations.

What Comes Next?

What happens next is still uncertain. Nvidia’s ability to sell H200 chips in China depends not just on US export licenses, but also on Chinese willingness to accept shipments and on evolving geopolitical priorities in both capitals.

If Nvidia begins shipments and generates meaningful revenue from China, it could signal a tentative thaw in tech trade tensions and provide a template for other companies seeking access to restricted markets. Conversely, if export control mechanisms stall or political blowback intensifies, Nvidia may find itself in a prolonged stalemate.

For now, the company’s cautious rollout — combined with continued enforcement scrutiny — shows how technology, trade and geopolitics are deeply intertwined in the race for AI supremacy, and how even small regulatory shifts can have outsized implications for global markets.