Los AngelesNew YorkChicagoHoustonPhoenixPhiladelphiaSan FranciscoSeattleAtlantaBoston
Tech · AI

32 Blackwell Ultra Chips Discovered in Chinese State Telecom Network

A security audit revealed that 32 export-restricted Nvidia processors were operating within the infrastructure of a major Chinese state-owned carrier.

By Robert Hayes ·

Thirty-two Nvidia Blackwell Ultra B300 graphics processing units have been identified within the network of China Telecom. This state-owned telecommunications provider serves a vast domestic market, and the presence of these specific chips raises immediate questions regarding recent U.S. export controls. The B300 model is currently barred from export to China, making its appearance in a Chinese data center a significant breach of trade regulations.

The discovery was made by Lava, an Israeli cybersecurity firm. The company reported that the chips had already been restricted by U.S. authorities. Under current rules, such high-performance hardware can only be sold to buyers on an approved list who agree not to resell or transfer the units to sanctioned parties. The finding suggests that the regulatory framework intended to limit China’s access to advanced AI compute is being tested.

This incident is not an isolated event in the ongoing struggle over semiconductor trade. Reports from 2025 indicated that Chinese entities smuggled one billion dollars worth of Nvidia AI chips into the country in just three months. At that time, the B200 series was the primary target, but distributors had already prepared to move the newer B300 series once it became available. The current discovery confirms that the flow of controlled hardware continues despite strict bans.

Nvidia has responded to the news with a measured tone. The company did not deny the findings but characterized the scale of the issue as small relative to the total market. A spokesperson stated that recent court cases have shown an insignificant diversion of controlled U.S. products, with figures often below half of one percent. The company described this amount as a drop in the bucket compared to the vast amount of domestic compute capacity China already possesses.

The spokesperson added that with tens of millions of GPUs in service in approved countries, the report demonstrates that smuggling is a non-factor. Nvidia noted that governments on both sides of the globe are actively blocking these transfers. The firm previously confirmed it is developing a tool to help locate its chips, though this solution has not yet been released to the public.

Lava identified the specific units using Nvidia’s own DCGM Exporter tool. This software is designed for data center operators to monitor server health and performance. While intended for internal use, the tool can sometimes be accessed from outside the local network. In some instances, it does not require password authentication, which may have facilitated the external discovery of the hardware. It remains unclear exactly where within China Telecom’s network the chips are located or how they were installed.

The demand for the B300 is driven by its superior specifications. Announced in March of last year, the chip offers 50 percent more memory and FP4 compute capability than its predecessor, the B200. Nvidia marketed the B300 as built for the age of reasoning, citing its suitability for sophisticated large language models. The chip was expected to ship by the end of 2025 and quickly became a highly sought-after asset in the global market.

Recent enforcement actions highlight the complexity of the smuggling networks. In August 2026, nine individuals in Taiwan were indicted for the illegal export of B300 GPUs to China. This followed an investigation by Supermicro, which fired several employees related to the smuggling. The network included a Taiwanese server trading firm called Flying Tiget Tech, which held whitelisted status for GPU purchases.

The high value of these components has also impacted the secondary market. Prices for older A100 servers, which are five years old, have tripled and can reach up to 82,000 dollars. The scarcity of new hardware has even led some users to retool gaming GPUs to handle AI workloads where possible. This shift underscores the critical role Nvidia’s silicon plays in modern data center operations.

Source: Tom's Hardware

More Tech · Report a business

Get Hot Press Today in your inboxCity news, clearly told. Free, a few times a week, unsubscribe any time.

Get Hot Press Today in your inboxCity news, clearly told. Free, a few times a week, unsubscribe any time.