The U.S. Commerce Department approved a little-known Chinese company to handle Nvidia’s H200 artificial intelligence chips despite evidence that the business has no functioning office at its registered address, according to a November 7 investigation by The New York Times. The findings have prompted new questions about how federal regulators examine foreign customers seeking access to advanced American semiconductor technology.

The company, New Matrix, received authorization to assemble H200 processors into AI computing systems under the Trump administration’s revised export licensing framework. However, reporters visiting its registered address in an industrial park on the northwestern outskirts of Beijing found an unfinished, apparently unoccupied office. The business was absent from the building directory, and property management representatives said they had no record of the company.

The New York Times investigation also identified connections between New Matrix’s owners and a Chinese private equity investment firm subject to U.S. trade restrictions. Those connections have raised concerns that ostensibly independent businesses could provide indirect access to sensitive American technology for companies facing export restrictions. The investigation did not establish that New Matrix unlawfully obtained processors or supplied chips to a prohibited recipient.

New Matrix was created approximately 17 months before the report, according to the newspaper. Publicly available information about its operations was limited, making it difficult to independently establish the scale of its business activities, staffing or technical capabilities. Although an empty registered office does not conclusively demonstrate that a company lacks legitimate operations elsewhere, the discovery raises questions about the information examined before its authorization was granted.

The corporate relationships identified in the investigation are particularly significant because of the U.S. government’s restrictions on Chinese businesses connected to sensitive semiconductor manufacturing and military technology. New Matrix’s owners have ties to Wise Road Capital, an investment firm that the Commerce Department added to its Entity List in December 2024. Two analysts who examined the corporate records reportedly assessed that New Matrix might be operating on behalf of companies associated with that investment group.

That assessment remains an allegation based on corporate relationships rather than an established finding of unlawful conduct. The available reporting does not identify New Matrix’s ultimate operational controller or demonstrate that restricted organizations received H200 chips through the company. A connection to a blacklisted business also does not automatically establish that every associated company is independently subject to the same restrictions.

The investigation extended beyond New Matrix. Another Chinese company approved to handle Nvidia processors, Maginfra, reportedly had an apparently deserted registered office when journalists visited its address. Corporate information also indicated connections between Maginfra and Inspur, a Chinese technology company placed under U.S. export restrictions because of concerns involving military modernization. Analysts cited in the reporting believed Maginfra could be an associated business, although that relationship does not independently establish an export violation.

The New York Times also examined corporate connections involving Zhejiang Yindun Cloud Technology and Shanghai Jiatang, two additional companies associated through their ownership or business relationships with organizations serving Chinese military-related customers. The reported approvals were part of a wider group of roughly two dozen Chinese companies authorized to purchase or handle H200 processors, including established businesses such as Baidu, BYD and Tesla’s Shanghai operations.

The findings put renewed attention on the Commerce Department’s Bureau of Industry and Security, which oversees export licensing and enforcement. According to accounts of the investigation, the agency said it carefully reviews applications for national security risks. It also pointed to enforcement activity, including approximately $324 million in penalties for export violations in 2025, as evidence of its regulatory efforts.

The current licensing arrangement resulted from a major policy change announced by President Donald Trump in December 2025. On January 13, 2026, the Commerce Department formally revised its review policy for exports of Nvidia H200 processors, AMD MI325X accelerators and comparable products to China. Instead of applying a general presumption of denial to qualifying transactions, regulators introduced case-by-case reviews subject to specified security requirements.

Under the revised policy, applicants must demonstrate that exports will not diminish global semiconductor manufacturing capacity available to American customers. Chinese purchasers must also maintain compliance procedures, including customer screening, while qualifying processors must undergo independent testing in the United States to verify performance and security. These conditions are intended to permit controlled commercial access while preserving safeguards against unauthorized diversion.

The New Matrix findings raise questions about whether licensing reviews are sufficient to identify the actual beneficiaries of individual transactions. Screening a registered company requires more than confirming its legal existence. Determining the ultimate recipient may also depend on understanding beneficial ownership, financial relationships, related entities and the physical facilities where equipment will be assembled or operated. The investigation highlights the difficulty of verifying those details across borders.

Nvidia’s H200 is a high-performance graphics processing unit designed for demanding artificial intelligence and scientific computing applications. Although Nvidia has introduced newer computing platforms, the H200 remains commercially valuable because of its processing capabilities, memory performance and compatibility with the company’s software ecosystem. Large numbers of these accelerators can support demanding AI training and inference workloads, making their international distribution a sensitive technology policy issue.

The licensing controversy comes amid complicated commercial conditions for Nvidia in China. Reuters reported on November 5 that approximately 600,000 H200 chips were being held in Hong Kong while awaiting authorization for shipment to mainland customers. The report cited people familiar with the situation who described Chinese government reluctance to permit large-scale imports of American processors, even as domestic technology companies continued seeking additional computing capacity.

According to Reuters, Chinese authorities had permitted limited H200 shipments into the mainland, with several major internet companies obtaining approximately 10,000 processors each in recent months. Some quantities reportedly moved through intermediaries and indirect shipping arrangements. Those broader transactions should not be confused with the New Matrix investigation: the available reports do not establish that New Matrix participated in the shipments described by Reuters.

The commercial restrictions have already affected Nvidia’s business. In its quarterly filing for the period ended July 26, 2026, the company stated that U.S. regulators had granted licenses for limited H200 deliveries to designated China-based customers. However, restrictions imposed by Chinese authorities prevented Nvidia from completing all authorized sales. The company reported a $400 million charge associated with excess H200 inventory and purchase obligations during the first half of fiscal 2027.

Nvidia also disclosed that shipments under the H200 licensing program accounted for less than 1% of its data center revenue in the most recently reported quarter. The filing said processors shipped through the program must undergo inspection in the United States before delivery and are subject to a 25% tariff upon importation into the country. Nvidia indicated that it had been unable to pass that tariff expense to customers.

The limited contribution from H200 sales to China illustrates the gap between the technology’s strategic significance and its immediate financial contribution to Nvidia. The company has substantial demand from American technology businesses investing in AI infrastructure, but China remains an important potential market. Export regulations, Chinese industrial policy and restrictions affecting particular customers all influence how much of that demand can translate into completed sales.

Nvidia has defended its broader approach to technology exports. In comments reported in connection with the investigation, company spokesman John Rizzo argued that reliance on American technology would present risks for the Chinese military because the United States could restrict access. He also emphasized Nvidia’s commitment to supplying advanced technology to American customers and pointed to Beijing’s efforts to encourage the use of domestically developed processors.

For U.S. regulators, the emerging question is not simply whether particular companies possess valid export licenses. It is whether the licensing system can identify the parties ultimately receiving or controlling advanced computing equipment after it leaves the original supplier. Complex ownership structures, contract manufacturing arrangements and multiple intermediaries can complicate that assessment, particularly when businesses have limited publicly observable operations.

The New Matrix investigation does not establish that American export rules were violated or that processors were transferred to prohibited military users. It does, however, identify circumstances that could warrant additional examination of corporate ownership, end-user representations and the effectiveness of existing compliance reviews. Any determination that specific companies breached export regulations would require additional evidence and an appropriate regulatory or legal process.

For Silicon Valley’s semiconductor industry, the consequences extend beyond Nvidia. American chip designers, hardware suppliers and data center technology companies operate within a regulatory environment shaped increasingly by national security considerations. Compliance requirements may affect customer relationships, equipment distribution, manufacturing partnerships and future international sales, especially where ownership structures are difficult to verify.

The November 7 findings leave several material questions unresolved, including where New Matrix conducts any actual assembly operations, whether the company has received H200 processors, and what information regulators considered before issuing authorization. Those questions will determine whether the case ultimately reflects an incomplete public picture of a legitimate business or a more substantial weakness in the screening of sensitive technology exports.

For now, the verified development is the disclosure of questionable corporate and physical-office information involving companies authorized to handle advanced Nvidia processors. The broader policy issue is whether Washington’s controlled-export framework can reliably distinguish legitimate commercial buyers from intermediaries potentially connected to restricted organizations, without preventing transactions that meet U.S. security requirements.