Blacklisted Chinese university bought and rented Nvidia H200 access | Magica
Blacklisted Chinese university bought Nvidia H200 chips and rented computing
Editorial Team
••📖7 min read
Northwestern Polytechnical University bought two Nvidia H200 processors for 874,700 yuan and awarded five contracts for rented Nvidia computing, procurement records show. Separate US restrictions on the defense-linked institution remained in place despite January’s conditional opening for H200 exports to China.
Northwestern Polytechnical University bought two H200 processors through a Chinese supplier for 874,700 yuan, about $123,000.
Five rental contracts also specified access to Nvidia processors subject to US restrictions.
The records leave the chips’ origins, the rented servers’ locations and the applicable authorizations unresolved.
Northwestern Polytechnical University, a Chinese defense-focused university that researches military aircraft, missiles and satellites, bought two Nvidia H200 AI processors in October despite longstanding US export restrictions, according to a report based on procurement records. An October 21 notice named Shaanxi Bosuan Network Technology as the supplier to its management school in the 874,700-yuan transaction.
The purchase was only one route to computing power. The university also awarded five contracts this year to rent access to powerful Nvidia processors from Chinese companies. Together, the transactions put domestic hardware suppliers and computing-service providers at the center of the question of how a restricted institution obtains US technology.
The evidence has limits. Procurement awards identify suppliers, equipment and intended work; they do not reconstruct the export chain or independently verify delivery and operation of every system. The report says Chinese public procurement agencies must verify suppliers’ ability to fulfill orders before announcing winners and signing contracts.
HGX H200 product image supplied by Nvidia; the university’s purchased cards and rented equipment are not pictured. Source: Nvidia.
A defense university remained restricted after the H200 opening
Based in Xi’an, Northwestern Polytechnical is one of China’s “Seven Sons of National Defense,” a group of elite defense-focused universities. US authorities added it to the Entity List in 2001 over alleged involvement in missile proliferation and military research, then tightened its restrictions in 2019, the procurement-based report says.
The Entity List is a US export-control mechanism that imposes licensing requirements on transactions involving listed parties. In the January 2026 regulatory compilation, the university’s entry covers all items subject to US Export Administration Regulations and specifies a presumption of license denial. It also names several university research institutes. The rules cover exports, reexports and transfers within a country, so a domestic seller does not by itself remove the licensing question.
ByteDance is reportedly accelerating plans to combine its Spell project with Doubao. Existing travel bookings reveal a transaction-fee business, while the phone assistant's history shows how access to other apps can constrain execution.
Washington’s January 15 H200 rule changed the review policy for qualifying exports from the United States to China and Macau from presumptive denial to case-by-case consideration. It did not lift the university’s separate restrictions.
To qualify, applicants must certify that the transaction excludes prohibited end users and end uses. They must also document customer screening to prevent prohibited parties from receiving remote access. Other conditions require assurances about US supply and foundry capacity, plus independent testing in the United States. Reexports and transfers within China did not receive the same opening.
Those safeguards apply to exports under that licensing route. Without the chips’ export histories, the university’s rental awards alone cannot establish whether those particular safeguards were breached.
US permission is also only part of the supply picture. The November 11 procurement report cites earlier reporting that small quantities of H200 chips had reached China through indirect routes while Beijing withheld approval for large-scale imports. That broader finding does not establish the source of this university’s processors.
The October purchase specified two H200 chips. Each of two seven-month January rental contracts specified eight; the records do not establish whether the contracts used distinct physical servers. Source: Yahoo News Malaysia / Reuters.
Two purchased chips, larger rented configurations
Nvidia, the computing company behind the processors, introduced the H200 in November 2023 as an accelerator for generative AI and scientific computing. Its launch announcement specified 141GB of high-bandwidth memory and described deployments through server manufacturers and cloud providers. The technology was designed for both owned systems and cloud access; the enforcement question concerns who receives that access.
Two cybersecurity-school contracts in January, each for seven months
A server containing eight H200 chips under each contract
441,700 yuan and 441,000 yuan
AI-model fine-tuning and reinforcement-learning training
Another university institute’s April contract
Computing services requiring at least 24 A100 chips
375,000 yuan
Structural and fluid-dynamics simulations
The university also rented access to servers containing eight A100 chips and eight A800 chips. Those older models remain barred from export to Chinese entities, according to the report. H200 was the most advanced processor cited in this university’s records.
The contract amounts are not comparable hourly prices. The two H200 rentals specify a seven-month term, but the reporting does not establish billable chip-hours, utilization or equivalent service terms across the awards. Nor do the configurations establish ownership of the rented chips or that separate contracts used distinct physical servers.
A wider market, with a documented precedent
The investigation examined 26 Chinese procurement and corporate records from 2026 and found other universities and research institutions buying restricted Nvidia chips or renting services through domestic suppliers. Several had defense-industry or military-research ties.
Beihang University, another Seven Sons institution with aerospace and defense-industry ties and research for military aviation and space programs, bought four A800 processors for 489,000 yuan in April. Other purchases involved H100 processors at Shenzhen University and Wuhan University of Technology.
Rental procurement extended beyond Northwestern Polytechnical. July and September government announcements selected suppliers of H800- and A800-based computing services for the University of Chinese Academy of Sciences. One listed H800 access at 10 yuan per chip-hour. That is evidence of an offered rental rate for a different processor and institution, not a benchmark for the H200 contracts.
An April 2024 investigation had identified 10 Chinese entities obtaining restricted Nvidia chips in servers made by Super Micro Computer, Dell and Gigabyte through resellers. It could not determine whether sellers used inventories acquired before restrictions tightened. Nvidia told the investigators that the tenders specified products exported and widely available before the restrictions and did not demonstrate violations by its partners. The server makers said they complied with applicable laws or would investigate further.
That earlier company explanation does not account for the current H200 purchases. Neither investigation’s selected contracts measure China’s total restricted-chip supply or prove that the processors were used in military systems.
The next test is tracing supply and access
The findings arrive amid a dispute over whether wider chip sales strengthen US technological influence or enable military capabilities. Nvidia’s founder and chief executive, Jensen Huang, has argued for continued US sales to preserve American leadership and Chinese reliance on US technology, according to the procurement report. Some US lawmakers argue that powerful AI chips can accelerate weapons design, intelligence analysis and military modernization.
The same report describes bipartisan calls for chip tracking and greater congressional oversight. Three Republican congressional committee chairmen have requested a November 18 briefing with Huang over his reported contacts with an alleged Chinese intelligence operative. That scrutiny supplies political context; it does not establish a connection to the university’s procurement.
Northwestern Polytechnical, Nvidia and the US Commerce Department’s Bureau of Industry and Security, which administers export controls, did not respond to requests for comment for the procurement report. Contact details for Shaanxi Bosuan Network Technology could not be determined.
Resolving the enforcement question requires evidence the awards do not provide: who supplied the domestic vendors, when and through which jurisdictions the chips moved, what authorizations applied, and where and how computing providers enabled the university’s access. Those records would distinguish documented procurement from a demonstrated breach and identify which parties were responsible.
The Office for Students wants more AI-free assessment and better AI skills teaching, backed by existing quality rules. Its research raises doubts about unsupervised coursework, but universities say assessment reform is already under way.
Finland has recognised Tesla’s provisional Dutch approval after changes to branding and speed controls. Drivers remain responsible, and the national decision does not commit Finland to backing EU-wide authorisation.
Grok Bot can connect to internal MCP servers through Tailscale without exposing them to the public internet. User-based authorization preserves network controls, but administrators must configure how narrowly the assistant can reach into their systems.
Daylight’s Goldman Sachs Alternatives-led Series B backs security agents supported by human engineers. The young company reports rapid growth, but has not disclosed the economics behind its automation promise.
Primordial Labs raised $18 million to widen deployment of Anura, its military command software. Existing contracts and demonstrations give it a foothold, but integration counts do not establish sustained use or lower operating costs.
The North Dakota Republican says allegations involving a suspected Chinese intelligence operative make Huang vulnerable to coercion. Nvidia’s board is backing its founder while independent directors investigate, with investors facing both security and leadership risks.
Scott Bessent says Washington and Beijing are establishing a channel for serious AI incidents, following their September agreement. Its usefulness depends on unresolved reporting procedures and participation by the companies that detect model failures.
Coframe's $25 million Series A backs AI agents that build and test websites and ads. Its offering combines automation with human expertise and customer approvals, while its claimed gains lack comparable measurement detail.
Alibaba’s Qwen4-Max preview gives developers a fixed model to test through November 25. API access requires Token Plan, while claimed coding gains remain unquantified in the launch materials.
Railway estimates stolen customer keys generated about $740,000 in unauthorized AI spending after an October intrusion. Secrets from up to about 100,000 workspaces could have been decrypted, and the company says it will cover losses providers decline to refund.
Anthropic expects retail investors worldwide to receive up to 20% of its IPO, potentially $12.9 billion at the proposed upper price. Access through six brokerage channels depends on demand and broker rules, while a separate technology-company investment would help finance vast infrastructure commitments and founders and a special trust retain control.
Anthropic's seven founders plan equal IPO share sales worth about $250 million each at the proposed midpoint. Their sales make up 2.9% of the public offering, while the company seeks much larger financing under a structure that limits public investors' voting influence.
Nvidia’s independent directors retained Wachtell Lipton partner Joon H. Kim for an investigation into reports of a foreign intelligence effort targeting co-founder and CEO Jensen Huang. The board previously expressed confidence in Huang; the appointment provides no findings or timetable.
RRE Ventures led the funding for Graphon’s enterprise memory infrastructure. The startup is pitching connected, persistent context across different agents, but benchmark gains leave deployment costs and commercial scale unresolved.
Goldman Sachs, a lead bank on Anthropic’s IPO, reportedly forecasts $685 billion in annual revenue by 2030. The Claude developer has substantial cash reserves and rapidly growing sales, but rising compute costs, competitive pricing and delivery risks complicate the enterprise growth case.
Memories Upload pays contributors for accepted first-person video and sells annotated data to robotics developers. Its million-hour collection claim leaves open how much footage improves robot performance, while contributors surrender ownership without royalties.
OpenAI has acquired AI travel concierge Odessia and is bringing its founder, former Sonder chief Francis Davidson, into its travel effort. The standalone service has closed, leaving its future role alongside ChatGPT’s existing travel partners unclear.
Anthropic says expected advances from future models have triggered its research-acceleration threshold. An outside assessment finds faster progress but does not certify compliance with that threshold, while the voluntary policy leaves consequential decisions about safeguards and continued development with the company.
OpenAI’s free business-profile beta adds branded agents for eligible U.S. ecommerce and online service firms. The agents can also serve as paid-ad destinations, while OpenAI controls distribution and businesses bear responsibility for outputs.
