NVIDIA is working directly with the U.S. Commerce Department to tighten controls on where its advanced AI processors end up, targeting a gray market that has allowed banned chips to reach China through intermediaries in Taiwan, Singapore, Malaysia, and Japan. The effort marks an escalation in enforcement of U.S. export restrictions and reflects growing concern that sophisticated smuggling networks are circumventing official sanctions.

How Are Smugglers Getting Banned Chips Into China?

For months, Chinese companies have found ways around U.S. export controls by purchasing NVIDIA's high-performance graphics processing units (GPUs) through shell companies and resellers in neighboring countries. These intermediaries, sometimes called "pass-through" entities, repackage the chips and route them to mainland China, effectively laundering the technology past U.S. restrictions. Taiwanese authorities raided the offices of Super Micro and other companies in late June 2026 as part of investigations into these shipment networks.

The problem is significant enough that U.S. prosecutors initiated legal action against Super Micro in March 2026 specifically to target these repackaging operations. Despite the company's insistence that it was working with authorities to ensure legal distribution, the smuggling continued at scale, creating what amounts to a backdoor into China's AI sector.

What Steps Is NVIDIA Taking to Stop the Leakage?

NVIDIA's response goes beyond passive compliance. The company has expanded its customer screening process significantly, moving from basic vetting to intensive on-site verification. The new approach includes:

• Facility Visits: NVIDIA teams now conduct in-person inspections of customer locations to verify they are legitimate businesses with genuine operations.
• Direct Interviews: Company representatives conduct detailed interviews with customers to assess whether purchases align with stated business needs.
• Geographic Focus: Screening efforts concentrate on high-risk transit countries including Singapore, Malaysia, and Japan, where chips are most likely to be diverted.
• Government Coordination: NVIDIA is working directly with the U.S. Commerce Department to align its enforcement efforts with federal export control policy.

The intensity of these measures has produced measurable results. According to industry insiders quoted in the Financial Times, NVIDIA's expanded screening has cut the company's Asian customer base roughly in half. This dramatic reduction reflects how aggressively the company is now vetting buyers.

What Impact Is This Having on the Chinese AI Market?

The crackdown is creating a genuine shortage of NVIDIA chips in China. Industry sources tell the Financial Times that large sales of NVIDIA's H200 GPUs, the company's most advanced model designed to meet Chinese specifications, are now unlikely. This matters because Chinese AI companies have relied heavily on NVIDIA hardware to train and run their large language models (LLMs), which are AI systems trained on vast amounts of text data to understand and generate human language.

The Chinese government has responded by exploring limited exemptions for some domestic firms, though Beijing has simultaneously blocked local companies from purchasing even the watered-down H200 variant that NVIDIA designed specifically for the Chinese market. This contradiction reflects China's dual strategy: reduce reliance on Western technology while maintaining enough access to stay competitive in AI development.

Why Does This Matter for the Broader Tech Competition?

The enforcement effort signals a hardening of U.S. export control policy and a willingness to use private companies as enforcement partners. NVIDIA's cooperation with federal authorities represents a shift from passive compliance to active policing of its own supply chain. This approach could become a model for other semiconductor manufacturers facing similar pressures to prevent technology leakage to restricted countries.

For China, the tightening noose around chip access forces a strategic choice: accelerate domestic chip design and manufacturing, or negotiate with the U.S. for limited exemptions. Neither path is quick or painless. Domestic alternatives take years to develop, while exemptions require political concessions that Beijing may be unwilling to make.

The practical effect is a bifurcation of the global AI hardware market, with Western companies serving Western and allied nations, while China builds its own ecosystem. This separation, if it holds, could reshape which countries lead in AI development over the next five to ten years.