H200 Clearance Gap: The U.S. Approved 10 Chinese Firms — Not a Single Chip Has Been Delivered

The United States has cleared approximately 10 Chinese companies to purchase Nvidia’s H200 AI chip, but not a single delivery has been made, according to a Reuters exclusive confirmed by three sources familiar with the matter and corroborated by CNBC reporting on May 14. The companies approved for H200 purchases include Alibaba, Tencent, ByteDance, and JD.com, along with distributors Lenovo and Foxconn. Each approved buyer is permitted to purchase up to 75,000 chips under the U.S. licensing terms.

Lenovo confirmed on the record to Reuters that the company “is one of several companies approved to sell H200 in China as part of Nvidia’s export license.” Other named firms did not respond to requests for comment. The U.S. Commerce Department declined comment.

The clearance gap — licenses issued, deliveries at zero — reflects simultaneous blockage mechanisms on both sides. Chinese firms have pulled back after guidance from Beijing, according to one source. U.S. rules impose certification requirements that have not yet been satisfied. The result is a bilateral stall in which the regulatory architecture of the deal is complete but the commercial transaction has not moved.

The U.S. Side: What Clearance Actually Requires

U.S. export rules issued in January require Chinese buyers to demonstrate they have installed “sufficient security procedures” and would not use the chips for military purposes before delivery can proceed. Nvidia must also certify sufficient inventory in the United States as a condition of any transfer.

A separate structural complication concerns the revenue arrangement negotiated by the Trump administration: a condition under which the U.S. would receive 25% of the revenue from chip sales. U.S. law does not permit the direct imposition of export fees, so the arrangement requires the chips to physically transit U.S. territory before shipping to China. This routing requirement adds logistical friction and has, according to sources, prompted concern in Beijing over potential tampering or hidden vulnerabilities in the hardware — a concern that sources describe as primarily a response to the physical routing structure rather than a judgment on Nvidia’s own security practices.

The China Side: Strategic Hesitation and the Domestic Push

Beijing’s hesitation reflects a distinct strategic calculation. China’s domestic AI chip development — led by Huawei and supported by companies like DeepSeek, which has publicly emphasized its reliance on domestic chips — is framed by Chinese planners as a long-run strategic asset. Approving large-scale imports of H200 chips at this stage would direct capital and deployment plans toward U.S.-sourced infrastructure, potentially undercutting the domestic industry rationale that Beijing is seeking to establish.

A fourth source told Reuters that scrutiny within China has intensified following two State Council supply chain security regulations that have prompted a government-wide effort to identify and eliminate potential foreign dependencies in critical technology infrastructure. The combination of strategic industrial policy, the routing-related security concern, and the supply chain review process has produced what Commerce Secretary Howard Lutnick described to a Senate hearing last month: “The Chinese central government has not let them, as of yet, buy the chips, because they’re trying to keep their investment focused on their own domestic industry.”

The Stakes: Jensen Huang’s Beijing Trip

Nvidia CEO Jensen Huang joined the U.S. delegation to Beijing after an invitation from President Trump, who picked him up in Alaska en route to a summit with Chinese President Xi Jinping, according to one source. Huang was not initially listed as part of the delegation. His inclusion signals that the H200 delivery stall was elevated to the diplomatic track — a notable escalation of a commercial dispute into a head-of-state agenda item.

Huang previously estimated China’s AI market would be worth $50 billion this year. Before U.S. export controls tightened, Nvidia held approximately 95% of China’s advanced chip market and derived roughly 13% of its total revenue from China. Huang has stated that Nvidia’s share of AI accelerators in China has effectively fallen to zero under the current restrictions. The H200 clearance structure represented a potential partial recovery of that position — a path that remains blocked despite regulatory approval.

The Policy Division Inside Washington

The stall has not produced consensus within Washington. Hardliners who opposed H200 sales to China have framed the continued delay as a positive outcome, arguing that chip sales to China reduce the inventory available to U.S. firms and that any deal accelerating Chinese AI infrastructure development undermines long-run U.S. competitive positioning. The Trump administration has maintained that H200 sales — under the 25% revenue arrangement and security conditions — would create a durable export channel rather than a strategic concession.

The two positions represent structurally incompatible views on the primary mechanism at risk: the hardliner position sees chip supply to China as the threat; the administration position sees lost commercial position and constrained diplomatic leverage as the cost of continued restriction. The H200 clearance gap — zero deliveries despite issued licenses — is the operational manifestation of that unresolved internal disagreement playing out in implementation.

Operator Takeaway

The clearance gap reveals that export license issuance and commercial delivery are distinct stages with their own blocking conditions. For operators tracking Nvidia’s China revenue outlook: the license structure is in place, but the dual-blockage mechanism — U.S. certification requirements plus Beijing’s strategic hesitation — means delivery timelines remain dependent on diplomatic resolution rather than commercial or regulatory process alone. The Huang-Trump-Xi Beijing interaction is the variable that could move the stall; a lack of outcome from that meeting would confirm that the gap is likely to persist through the remainder of 2026. China’s domestic chip push, reinforced by the State Council supply chain directives, creates a structural incentive for Beijing to delay regardless of U.S.-side conditions — a factor that persists independent of diplomatic outcome.

FAQ

What is the H200 chip and why does it matter for China?

The H200 is Nvidia’s second-most powerful AI chip, below the H100/B100 series that remains fully restricted for China export. Before U.S. export controls tightened, Nvidia held approximately 95% of China’s advanced AI chip market. The H200 export clearance represented a partial reopening of that market under specific conditions. China’s AI infrastructure — including large language model training and inference at scale — requires high-performance GPUs that Nvidia’s domestic competitors, primarily Huawei, have not yet matched in volume or performance.

Which Chinese companies were cleared to buy H200 chips?

According to Reuters sources, approximately 10 firms received clearance. Named companies include Alibaba, Tencent, ByteDance, and JD.com as direct buyers, and Lenovo and Foxconn as authorized distributors. Lenovo confirmed its distributor status on the record. Each approved buyer is permitted to purchase up to 75,000 chips under the licensing terms.

Why have no chips been delivered despite U.S. approval?

Two simultaneous blocking mechanisms apply. On the U.S. side, buyers must certify sufficient security procedures and non-military use, and Nvidia must certify U.S. inventory levels. The revenue arrangement also requires chips to transit U.S. territory, adding logistical and security-perception friction. On the China side, firms reportedly pulled back after guidance from Beijing, which is prioritizing domestic chip development and conducting a supply chain security review that discourages new dependencies on foreign-sourced AI infrastructure.

What is the 25% revenue arrangement?

The Trump administration negotiated a condition requiring the U.S. to receive 25% of revenue from H200 chip sales to China. Because U.S. law does not permit direct export fees, the arrangement requires chips to physically pass through U.S. territory before delivery to Chinese buyers. Sources describe this as primarily a legal workaround rather than a security measure, but it has generated concern in Beijing about potential hardware tampering or surveillance vulnerabilities.

Why did Jensen Huang join the Beijing delegation?

Huang was added to the U.S. delegation to Beijing after an invitation from President Trump, who picked him up in Alaska en route to meetings with Chinese President Xi Jinping. Huang was not initially listed as part of the delegation. His inclusion suggests the H200 delivery stall was elevated to the diplomatic agenda — an indication that the commercial gap has reached a level of strategic significance requiring direct engagement at the head-of-state level.