Export Control Tensions: B300 Chips Discovered in China Telecom Network
Security firm identifies prohibited Nvidia Blackwell Ultra processors in a major Chinese state network, prompting Nvidia to dismiss the finding as a minor anomaly.
By News Media House Staff ·

The integrity of U.S. export controls faces renewed scrutiny after an Israeli security firm identified thirty-two Nvidia Blackwell Ultra B300 chips operating within the network of China Telecom. This state-owned telecommunications provider serves the domestic Chinese market, making the presence of these specific processors a significant development in the ongoing struggle to restrict high-performance artificial intelligence hardware. The discovery raises immediate questions about the efficacy of current trade restrictions and the methods used by various actors to bypass them.
The B300 chips are part of a class of high-performance graphics processing units that the United States has barred from direct export to China. Under current regulations, these devices can only be sold to approved buyers who agree not to transfer them to sanctioned parties. The finding by Lava, the security company behind the report, highlights a persistent gap between policy and practice in the global semiconductor supply chain.
This incident follows a broader trend of chip smuggling. In 2025, reports indicated that Chinese entities moved one billion dollars worth of Nvidia AI chips into the country over a three-month period. At that time, the B200 was the primary target of distributors. Now, with the B300 series available, the flow of restricted technology continues, suggesting that demand for advanced compute power remains robust despite regulatory hurdles.
Nvidia has responded to the findings with relative calm. A company spokesperson stated that governments, researchers, and media have identified insignificant diversion of controlled U.S. products, noting that recent court cases showed figures less than one half of one percent. The spokesperson described this volume as a drop in the bucket compared to the ocean of domestic compute China already possesses. With tens of millions of GPUs in service in approved countries, Nvidia maintained that smuggling remains a non-factor, largely blocked by governments on both sides of the globe.
The mechanism of detection offers insight into the vulnerability of current data center monitoring tools. Lava identified the chips using Nvidia’s DCGM Exporter tool. This software is intended for internal use by data center operators to monitor server health and performance. However, it is often accessible from beyond the internal network and does not always require password authentication. Nvidia has acknowledged working on a tool to locate its chips but has yet to release it.
The Blackwell Ultra B300 is a highly sought-after piece of hardware. Announced in March of last year, the processor offers substantial performance gains, including fifty percent more memory and FP4 compute capability than its predecessor, the B200. Nvidia positioned the chip as being built for the age of reasoning, citing sophisticated large language models like DeepSeek R1. The chip was expected to ship by the end of 2025 and quickly became a target for illicit trade.
The smuggling pipeline is complex and often involves intermediaries. In August 2026, nine individuals were indicted in Taiwan for the illegal export of B300 GPUs to China. This legal action occurred days after Supermicro confirmed it had fired several employees following an internal investigation. The scheme reportedly involved a Taiwanese server trading firm with whitelisted status for GPU purchases, highlighting the role of third-party vendors in the chain of custody.
The high demand for these chips has driven up the price of older inventory. Nvidia’s A100 servers, now five years old, have seen their prices triple, reaching as much as eighty-two thousand dollars. The scarcity of new, high-end chips has also led some users to retool gaming GPUs for AI workloads, a testament to the critical role these processors play in the current AI infrastructure landscape.
Source: Tom's Hardware


