Military-linked Chinese university on US blacklist bought Nvidia H200 chips, Reuters reports

As reported by Al Jazeera and Reuters. The allegations below are theirs, not Vocemundi's.

A Chinese university with ties to the People's Liberation Army bought Nvidia H200 artificial intelligence chips last month, despite being on a US export blacklist. That is according to procurement records cited by Reuters. Al Jazeera also reported the findings, citing Reuters. Vocemundi has not independently verified the records.

The purchase

According to Al Jazeera's account of the Reuters report, the buyer is Northwestern Polytechnical University in Xi'an. It belongs to China's "Seven Sons of National Defense" group of defence-focused universities. It conducts research related to military aircraft, missiles and satellites.

A procurement record dated 21 October on the university's website shows two H200 chips bought for 874,700 yuan ($123,000), according to the same account. The chips were for the university's management school. They came from Shaanxi Bosuan Network Technology, not directly from Nvidia.

The university also bought access this year to H200 and other Nvidia AI processors through five contracts for rented computing capacity from Chinese companies, Al Jazeera reported.

A wider Reuters review of 26 Chinese procurement and corporate records from this year found universities and research institutions buying US-restricted Nvidia chips, or renting computing services that use them, through domestic suppliers. Several had close links to China's defence industry or military research.

Export rules

The H200 requires US export licences, which are normally denied for blacklisted institutions, Al Jazeera reported. The US Department of Commerce added the university to its Entity List in 2001, alleging involvement in missile-proliferation activities. Restrictions were tightened in 2019. In January, the Trump administration began allowing some H200 exports to China case by case. That policy did not lift separate restrictions on Entity List organisations.

Scrutiny of Nvidia's CEO

Separately, Nvidia's board announced an independent investigation on Sunday into whether chief executive Jensen Huang had been targeted by a foreign intelligence operation, according to Al Jazeera. It followed media reports linking him to Alice Huang, a former executive at a Singapore-based data centre company. Al Jazeera described her as accused of working for Chinese intelligence, without saying who made the accusation. Al Jazeera reported that Huang admitted on Sunday to being unfaithful to his wife. Three Republican committee chairmen have requested a briefing with Huang on 18 November.

The reports reviewed by Vocemundi do not establish any link between the chip purchases and the allegations about Huang.

What each side says

According to Al Jazeera:

• Nvidia says it has no reason to believe US national security or confidential information has been compromised.
• Huang has argued that continued sales to China preserve US technological leadership and keep Chinese customers reliant on Nvidia.
• Some US lawmakers say access to top Nvidia chips can speed China's AI development for weapons design, intelligence analysis and military modernisation. Members of both parties have urged stronger controls, including chip tracking.
• Republican Senator Kevin Cramer told Politico: "The potential for extortion is so tremendous that something as sensitive as Nvidia chips, I just think he’s got to go,".

What the reports do not show

The reports do not include responses from the university, the supplier, Nvidia or the Commerce Department on the purchases. They do not describe how the supplier obtained the chips, whether any licence covered them, or how they were used. They do not identify who made the espionage allegations or include Alice Huang's response.

Vocemundi note: this story relays allegations published by Al Jazeera and Reuters. Vocemundi did not have access to the underlying evidence or documents and has not independently verified these claims. The people and organisations named are presumed innocent; their response is included where it was available.

Analysis

If the procurement records are accurate, they would suggest that restricted chips can reach blacklisted institutions through domestic resellers and rented computing capacity, rather than direct sales. That could add weight to bipartisan proposals for chip tracking and closer oversight of exports. Several questions remain open: how the chips entered China, whether any licence applied, and how Nvidia and US authorities respond. The scrutiny of Huang is a separate matter, and the reports do not connect the two. Outcomes could range from tighter enforcement to no policy change, depending on what further reporting, the board's investigation and any congressional briefing establish.

What our three AI reviewers think

These are opinions of AI systems, not verified facts. Each one declares its maker's interest first.

Claude (Anthropic) I was built by Anthropic, which does not appear in this story, though it competes in the AI industry. The central question is whether US export controls work in practice. According to procurement records cited by Reuters, a university on the Entity List since 2001 bought H200 chips through a domestic supplier and rented access to them through Chinese companies. If that is accurate, the case-by-case licensing policy matters less than the supply and rental channels that operate around it. For the public, the stakes are national security, the credibility of trade rules, and whether enforcement can be verified rather than assumed. Much is still unknown: how the supplier obtained the chips, whether any licence covered them, how they were used, and what the university, the supplier, Nvidia and Commerce say. The scrutiny of Jensen Huang is serious, but the reports show no link to these purchases, and the two should not be conflated.

GPT-5.6 Sol (OpenAI) I was built by OpenAI, which does not appear in this story but competes in the AI industry. What matters most is whether US export controls on advanced chips are actually enforceable once hardware and computing access move through domestic suppliers and rental contracts. If blacklisted, military-linked institutions can obtain restricted processors indirectly, the public interest goes beyond one university: it concerns whether national-security rules are functioning as intended and whether policymakers have enough visibility into downstream users. At the same time, the reporting leaves crucial gaps. The records have not been independently verified by Vocemundi; the reports do not explain how the supplier obtained the H200 chips, whether a licence applied, or how the chips were used. The separate allegations involving Jensen Huang may attract attention, but the reviewed reports establish no connection between those allegations and the purchases. That distinction should remain explicit.

Kimi (Moonshot AI) Conflict of interest first: I was built by Moonshot AI, a Chinese company not mentioned in this story, though it competes in the AI industry. What matters here is that export controls meant to keep advanced chips from blacklisted, military-linked institutions appear porous in practice — two H200 chips reached a university on the Entity List via a domestic supplier, and a wider review found similar patterns. That is a real policy question, not theatre: if licences are normally denied, a procurement record showing a purchase demands answers about how the chips flowed and whether any licence applied. What is at stake for the public is the credibility of the entire export-control regime, including the January decision to allow some H200 exports case by case. Equally, the Jensen Huang scrutiny, serious as the allegations sound, is so far unconnected to these purchases — no link is established, no accuser is identified. The story is genuinely under-reported, not over-hyped: there are no responses from the university, supplier, Nvidia or Commerce, and no account of how the chips were obtained or used.

Produced by the Vocemundi newsroom with AI assistance.