Urgent.News

What's breaking now, across thousands of outlets.

AI

Two Chinese companies cleared to buy banned Nvidia AI chips have empty offices and unknown addresses

The offices for New Matrix and Maginfra, both approved to receive H200 chips, appear abandoned, and people in the buildings say they're unfamiliar with the companies or have never seen anyone working there.

Two Chinese companies cleared to buy banned Nvidia AI chips have empty offices and unknown addresses

Two Chinese firms have been authorized to acquire Nvidia H200 AI chips, yet their corporate headquarters appear to be unoccupied, with receptionists demonstrating a lack of familiarity with these firms. According to The New York Times, these entities have ties to other enterprises sanctioned by the United States for cooperation with the Chinese military.

Former White House National Security Council China director Liza Tobin expressed skepticism over both the U.S. Department of Commerce and Nvidia conducting adequate inspections to ensure compliance with export licensing terms. Among the entities granted approval are well-known names such as Baidu, BYD, Tesla Shanghai, ZTE, Alibaba, Tencent, and ByteDance.

However, the investigation also identified two lesser-known companies: New Matrix and Maginfra. Both firms operate from vacant offices with minimal furnishings, and nearby workers have never heard of them. New Matrix's website reveals no information on its owners or history, while public records disclose that it was established by Li Bin, who shares connections with a Chinese private equity firm sanctioned by the U.S. for facilitating the acquisition of sensitive chip technology.

Li also chairs Tsinghua Unigroup, a tech conglomerate with governmental ties, and owns New H3C Semiconductor Technologies, which was added to the Commerce Department's Entity List for partnering with the Chinese military. Maginfra is suspected of being associated with Inspur, a firm blacklisted by Washington in 2023 due to its support for the People's Liberation Army, sharing personnel, patents, and a similar government ownership structure.

These two entities are not the ultimate recipients of Nvidia's AI chips but are entrusted with assembling them into servers, which will eventually reach a final user, but their identities remain unclear. Another company, Zhejiang Yindun Cloud Technology, has executives linked to the founder of Range Intelligent Computing Technology Group, a data center operator that supplied compute capacity to Chinese military contractors, and operates partially from a Range data center.

The initial export ban imposed by President Donald Trump in April 2025, which included customized Nvidia AI GPUs, was lifted in exchange for a 25% reduction in the sale, but the first H200 shipments only began arriving in China over eight months later.

Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at tomshardware.com →

More in AI

More from Sunday 8 November →