China-Linked University Obtains Nvidia H200 AI Chips Despite US Blacklist
China's access to advanced artificial intelligence hardware is raising fresh questions about the effectiveness of American technology export restrictions. Procurement records reveal that a Chinese university with close military connections acquired Nvidia's powerful H200 processors through domestic suppliers.
A new development in the US-China technology rivalry has brought the global semiconductor industry back into focus. Northwestern Polytechnical University, a Chinese institution with longstanding connections to the country's defence sector, purchased two Nvidia H200 artificial intelligence processors in October 2026 despite being listed under US export restrictions.
The purchase highlights the growing challenges facing Washington as it attempts to restrict China's access to advanced computing technology. It also raises questions about how effectively export controls can be enforced when restricted hardware and computing services are available through third-party suppliers.
Two Nvidia H200 Chips Purchased for $123,000
Northwestern Polytechnical University purchased two Nvidia H200 processors for 874,700 yuan, approximately $123,000, according to procurement records dated October 21.
The chips were acquired for the university's management school through a Chinese technology supplier, Shaanxi Bosuan Network Technology, rather than directly from Nvidia.
The H200 is among Nvidia's most powerful AI processors. It is designed to support demanding artificial intelligence workloads, including training and running advanced AI models.
The purchase is particularly significant because Northwestern Polytechnical University is included on the US Commerce Department's Entity List. Organisations on this list face restrictions on obtaining certain American technologies without government authorisation.
The transaction therefore raises questions about how a university subject to strict export restrictions was able to secure access to such advanced hardware through a domestic supplier.
Why the University Is Under US Restrictions
Northwestern Polytechnical University is based in Xi'an, China, and is known for research involving aerospace engineering, military aircraft, missiles and satellite technologies.
The institution belongs to China's group of defence-focused universities commonly known as the Seven Sons of National Defence.
The US government placed the university on its Entity List in 2001 over concerns involving alleged missile-proliferation activities and military research. Restrictions were strengthened in 2019, with export licences for controlled American technologies generally subject to a policy of denial.
These restrictions are intended to limit the availability of sensitive American technology to organisations that Washington considers potential national security risks.
Access to advanced AI processors is particularly important because artificial intelligence can support a wide range of applications, from scientific research and industrial automation to defence-related computing.
More Than Just a Direct Chip Purchase
The two H200 processors were not the university's only reported access to Nvidia technology.
Procurement records also show that Northwestern Polytechnical University entered into five agreements during 2026 to rent computing capacity from Chinese technology companies.
Two seven-month agreements awarded by its cybersecurity school in January involved access to servers containing eight H200 processors each. The contracts were valued at 441,700 yuan and 441,000 yuan.
The services were intended for activities including AI model fine-tuning and reinforcement-learning training.
The university also obtained access to computing resources containing Nvidia A100 and A800 processors.
Another contract, awarded in April by a university institute, involved computing services requiring at least 24 A100 chips for structural and fluid-dynamics simulations.
This distinction between purchasing physical processors and renting computing capacity is important. An organisation does not necessarily need to own restricted hardware to benefit from its computing power. Access to servers operated by third-party companies can provide similar capabilities for specific research and development tasks.
Other Chinese Universities Also Obtained Restricted Chips
The procurement records point to a wider pattern involving Chinese universities and research institutions.
Shenzhen University purchased three Nvidia H100 processors for 769,800 yuan in September. Wuhan University of Technology acquired a server containing an H100 processor for 355,000 yuan and separately purchased four A100 processors for 344,000 yuan.
Beihang University, another institution belonging to the Seven Sons of National Defence group, purchased four Nvidia A800 processors for 489,000 yuan in April.
Other procurement announcements also identified computing services involving Nvidia H800 and A800 processors for the University of Chinese Academy of Sciences.
A review of 26 procurement and corporate records identified purchases of restricted Nvidia processors or access to computing services using such hardware.
However, the available records do not establish the original supply routes for every processor or the physical locations of all the rented servers.
This uncertainty makes it difficult to determine the complete scale of restricted chip access in China.
Why US Export Controls Are Facing Challenges
The latest findings highlight an important weakness in technology export restrictions: controlling direct sales does not necessarily prevent access through complicated supply chains.
US export regulations generally require authorisation for shipments of controlled processors to restricted organisations. However, domestic suppliers, resellers and computing-service providers can create additional layers between the original manufacturer and the final user.
This makes monitoring more complicated, particularly when customers rent access to computing infrastructure instead of purchasing processors directly.
The Trump administration began allowing certain H200 exports to China on a case-by-case basis in January 2026. However, this policy did not remove the separate restrictions applicable to organisations such as Northwestern Polytechnical University.
The licensing framework also includes safeguards intended to prevent blacklisted organisations and military end-users from remotely accessing exported processors.
The procurement findings raise questions about how these safeguards can be monitored effectively.
Nvidia and the Bigger US-China AI Competition
The developments also place Nvidia at the centre of an increasingly complicated geopolitical debate.
Nvidia CEO Jensen Huang has argued that allowing American companies to continue selling technology to China would help maintain US technological leadership while keeping Chinese customers connected to American computing platforms.
However, some US lawmakers have expressed concerns that advanced AI processors could accelerate China's military research and technological modernisation.
They have called for stronger export monitoring, chip-tracking mechanisms and greater congressional oversight.
For Nvidia, China represents an important technology market, but government restrictions create uncertainty about how much of that demand the company can legally serve.
For Washington, the challenge is balancing the commercial interests of American semiconductor companies against national security priorities.
What This Means for the Global AI Industry
The latest developments demonstrate that access to advanced AI computing is becoming an increasingly important part of international competition.
High-performance processors are essential infrastructure for developing sophisticated AI systems. Restrictions on their availability can influence research capabilities, technology investments and the competitive position of companies and countries.
The reported purchases do not establish the complete extent of China's access to restricted Nvidia hardware. Nevertheless, they demonstrate why export enforcement requires more than controlling direct manufacturer-to-customer transactions.
Washington may face increasing pressure to strengthen oversight of resellers, cloud computing providers and international semiconductor supply chains.
The bigger question is no longer simply whether restricted AI chips can be sold to China. It is whether governments can effectively control who ultimately gets to use their computing power.
