Nvidia’s H200 Sales to China Put Export Screening Under Scrutiny

Introduction

The latest dispute over Nvidia’s AI-chip sales to China is less about the H200’s technical specifications than about the identity of the companies allowed to receive it. According to reporting by The New York Times, as relayed by The Verge AI, a Chinese company called New Matrix appeared on a list of firms approved to handle Nvidia’s H200 chips. Yet the company’s listed office reportedly showed few signs of an operating business.

The finding does not by itself prove that New Matrix received chips, violated export rules, or transferred hardware to a prohibited organization. It does, however, raise a basic question for an export-control system built around national-security concerns: how thoroughly are regulators and suppliers checking the real-world existence, ownership, and activities of approved customers?

Key points

• The company’s physical presence is difficult to verify. New Matrix was created relatively recently and has little visible information online. The building directory at its registered address reportedly did not list the company. Staff at a nearby property-management office also said they had no record of it. The space itself was described as largely empty and unfinished.
• Its ownership creates an additional compliance concern. The company’s owners have ties to a Chinese private equity firm that has been on a US blacklist since 2024. The available material does not establish that New Matrix is itself sanctioned or that it has breached export rules. The connection does make deeper ownership and control checks especially important.
• The report describes a wider pattern. The New York Times said the case was not necessarily an isolated oversight. Other companies linked to entities facing national-security-related trade restrictions have reportedly been approved to buy Nvidia chips. If confirmed, that would suggest a gap between formal list screening and meaningful due diligence.
• The policy context has shifted. Nvidia faced restrictions on selling certain advanced chips to China during the Biden administration. The report says the Trump administration later eased those limits after Nvidia agreed to give the US government 15 percent of sales. Chief executive Jensen Huang also mounted an intensive lobbying effort in Washington.

Why it matters

Export controls depend on identifying the actual end user, not merely checking whether a company name appears on a sanctions list. A newly formed company with limited staff, unclear operations, or opaque ownership can become a potential channel for rerouting sensitive hardware even when its legal name is not directly restricted. Effective screening therefore needs to examine the company’s premises, employees, financial capacity, beneficial owners, business purpose, and the eventual location and use of the chips.

The issue also creates risk for Nvidia. A relaxation of sales restrictions does not eliminate supplier responsibility. If customer reviews rely too heavily on registration documents or self-reported information, the company could face reputational damage, enforcement scrutiny, and renewed political pressure. For Washington, granting licenses and monitoring what happens after approval are separate tasks; one cannot substitute for the other.

The material available here does not include a full response from Nvidia or the Commerce Department, nor does it demonstrate that H200 chips reached New Matrix or a prohibited destination. The most defensible conclusion is therefore narrower: the reported case exposes serious questions about end-user verification, but the ultimate flow of hardware and any legal violations remain unconfirmed.

Source: The Verge AI