Chip Smuggling: 32 Banned Nvidia B300s Found in China Telecom's Network

Thirty-two Nvidia Blackwell Ultra B300 graphics processing units have been identified inside the infrastructure of China Telecom, a state-owned telecommunications giant in China. These chips are particularly significant because they are part of Nvidia’s latest high-performance line, which has been explicitly barred from export to China by U.S. regulations. The discovery was made by Lava, an Israeli cybersecurity firm, which noted that these specific processors are intended only for whitelisted buyers who agree not to resell them to sanctioned entities. The presence of these units in a major Chinese state network highlights the continuing difficulty in enforcing export controls on some of the most advanced silicon available today.

The timeline of how these chips ended up in Chinese server racks suggests a persistent gap between regulatory intent and market reality. In 2025, reports indicated that Chinese companies smuggled approximately $1 billion worth of Nvidia AI chips into the country in just three months. At that time, the B200 model was the primary target, but distributors were already positioning themselves for the B300 series as it became available. This latest finding confirms that the flow of restricted hardware has not stopped, even as the technology evolves.

Nvidia has responded to the report with a measured level of skepticism regarding its broader impact. While the company did not deny that the chips were found, a spokesperson characterized the incident as a minor anomaly in a much larger market. According to Nvidia, government and media investigations have typically found that the diversion of controlled U.S. products accounts for less than half of one percent of total sales. The company argued that compared to the "ocean of domestic compute" China already possesses, these smuggled units represent a "drop in the bucket." Nvidia emphasized that with tens of millions of GPUs already in service in approved countries, smuggling remains a non-factor that is effectively managed by governments on both sides of the Pacific.

The method by which Lava identified the chips offers a glimpse into the technical vulnerabilities of data center monitoring. The firm used Nvidia’s own DCGM Exporter tool to locate the B300 units. This software is designed for internal use by data center operators to monitor server health and performance. However, it can sometimes be accessed from outside the immediate network and, in many cases, does not require password authentication. While Nvidia has confirmed it is developing a dedicated tool to help locate its chips, that solution has not yet been released. It remains unclear exactly where within China Telecom’s vast network the 32 chips are housed, nor how they physically arrived there.

The demand for the Blackwell Ultra B300 is driven by its substantial performance advantages over its predecessor. Announced in March of last year, the chip offers 50% more memory and FP4 compute capacity than the B200. Nvidia positioned the B300 as hardware "built for the age of reasoning," specifically highlighting its suitability for sophisticated large language models like DeepSeek R1. Originally expected to ship by the end of 2025, the chip quickly became a high-value target for the black market.

Recent legal actions in Taiwan further illustrate the complexity of the smuggling routes. In August 2026, nine individuals were indicted in Taiwan for the illegal export of B300 GPUs to China. This development occurred days after Supermicro, a major server manufacturer, confirmed it had fired several employees following an internal investigation into smuggling activities. The operation reportedly involved a Taiwanese server trading firm known as Flying Tiget Tech, which held whitelisted status for purchasing GPUs, allowing it to act as a conduit for the restricted hardware.

The scarcity of high-end Nvidia chips has also triggered secondary effects in the broader market. The critical importance of these accelerators to AI data centers has driven up the prices of older silicon. Nvidia’s A100 servers, which are five years old, have seen their prices triple, with units now costing up to $82,000. In some cases, operators have resorted to retooling gaming GPUs to handle AI workloads where possible, a workaround that underscores the intense pressure on supply chains.

For readers in Texas and the South Central states, this story underscores the global race for computational power that underpins modern artificial intelligence. As U.S. export controls tighten, the methods used to bypass them become increasingly inventive. While Nvidia maintains that the volume of smuggled chips is manageable, the presence of 32 top-tier B300 units in a state-owned Chinese network serves as a reminder that the battle for AI dominance is fought not just in boardrooms, but in the physical logistics of server racks and shipping containers.

Looking ahead, the release of Nvidia’s new chip-location tool will be a key development to watch. Until then, security firms like Lava will likely continue to scan for anomalies in exposed monitoring tools. The interplay between U.S. regulations, Chinese demand, and the technical capabilities of data center operators will determine whether these "drops in the bucket" remain a nuisance or grow into a more significant challenge for the global tech landscape.