Blacklisted Chinese University Bought Four Nvidia H200 Cards
Reuters-reviewed procurement records show blacklisted Northwestern Polytechnical University bought four Nvidia H200 cards and rented restricted GPU servers through Chinese domestic suppliers.
By Daniel Okafor
4 min read

What's News
- Northwestern Polytechnical University bought four Nvidia H200 cards for a combined 1,670,700 yuan through domestic suppliers, Reuters reported on November 11, 2026.
- The university has been on the U.S. Entity List since May 14, 2001, with license requirements since September 20, 2016.
- Two January 28, 2026 awards covered rented HGX H200 eight-GPU systems totaling 882,700 yuan for AI fine-tuning.
- The January 15, 2026 U.S. rule moved H200 exports to case-by-case review but did not lift the university's Entity List restrictions.
- Shenzhen University bought three H100 processors for 769,800 yuan; Beihang University bought four A800 processors for 489,000 yuan.
Northwestern Polytechnical University bought four Nvidia H200 AI accelerator cards for a combined 1.67 million yuan despite sitting on the U.S. Commerce Department's Entity List since 2001. Reuters reported the purchases on November 11, 2026, after reviewing 26 Chinese procurement and corporate records showing universities buying U.S.-restricted Nvidia chips or renting Nvidia-powered computing through domestic suppliers.
The Xi'an-based university did not buy from Nvidia directly. A management school award notice dated October 21, 2026 lists two NVIDIA H200 141G PCIe data-center AI accelerator cards for 874,700 yuan, supplied by Shaanxi Bosuan Network Technology. A separate cybersecurity school notice dated September 4, 2026 lists two more H200 cards for 796,000 yuan. The H200 trains and runs advanced AI systems, and exporting it to China requires U.S. licensing approval.
The university is a poor candidate for such approvals. It researches military aircraft, missiles and satellites, and it belongs to China's Seven Sons of National Defense group of defense-focused universities.
How the university rented restricted computing
Beyond hardware purchases, the institution obtained computing power without owning the machines. Two awards from its cybersecurity school, both dated January 28, 2026, covered separate NVIDIA HGX H200 eight-GPU systems. The first was a 441,000-yuan reinforcement-learning fine-tuning service running from February 16, 2026 through September 16, 2026. The second was a 441,700-yuan multitask fine-tuning service running from February 1, 2026 through August 31, 2026. Both contracts specified 141GB of memory per GPU and supplied server capacity, not additional H200 cards.
Two further January 28, 2026 awards covered older chips barred from export to all Chinese entities. One covered a server with eight Nvidia A800 80GB GPUs for training-data processing at 240,000 yuan for 12 months. The other covered a server with eight A100 SXM4 80GB GPUs for model inference at 471,600 yuan for 12 months. In April 2026, another university institute agreed to pay 375,000 yuan for computing services requiring at least 24 A100 chips for structural and fluid-dynamics simulations.
Why the 2026 licensing change did not help
A January 2026 rule change for some H200 exports left the university's restrictions untouched. The Commerce Department added Northwestern Polytechnical University to the Entity List on May 14, 2001. Since September 20, 2016, a license has been required to export, reexport or transfer any item subject to the Export Administration Regulations to the university, according to a Bureau of Industry and Security enforcement order. Restrictions tightened again in 2019, making licenses for controlled U.S. technology exports to the university subject to denial in most circumstances.
The January 15, 2026 rule moved qualifying H200 and equivalent-chip exports to end users in China or Macau from a presumption of denial to case-by-case review. But that route carries its own conditions, and each one excludes a blacklisted defense university:
- Applicants must certify sufficient U.S. supply and no diversion of global foundry capacity that would otherwise serve U.S. customers.
- They must arrange independent third-party testing in the United States.
- They must document customer screening and physical security.
- They must confirm the commodities are not for military or military-intelligence end uses or users.
- They must confirm no transaction party is covered by the Entity List provision and that restricted parties will not receive remote access.
The revised route therefore does not override the university's Entity List restrictions.
What the wider records show
Other institutions appear in the same review. Shenzhen University bought three H100 processors for 769,800 yuan in September 2026. Beihang University bought four A800 processors for 489,000 yuan in April 2026. Beihang also belongs to the Seven Sons group and maintains close ties to China's aerospace and defense industries.
Chinese public-procurement rules require agencies to verify that suppliers can fulfill orders before announcing winning bidders and signing contracts. The records offered little information about where the processors originated or where the rented servers operated. They establish purchases and contractual access while leaving the chips' upstream supply chain and any specific export violation unresolved.
Why enforcement now looks domestic as well as border-focused
The disclosure lands in the middle of a policy fight in Washington. Nvidia chief executive Jensen Huang has said continued U.S. sales to China would preserve American technological leadership and keep Chinese customers reliant on U.S. technology. Lawmakers in both parties have urged stronger controls, including chip-tracking requirements and greater congressional oversight of exports.
The records reframe the enforcement problem. Restricting licensed exports does not by itself prevent a blacklisted institution from acquiring AI accelerators or renting their computing capacity through Chinese suppliers. The challenge now runs through the domestic Chinese market as much as the border.
Source: unite.ai
More from Daniel Okafor
Show full bio
Correspondent covering money, funding and market shifts around entrepreneurship, startups and venture funding at Business Bearings.
1,980 articles