Smuggled Nvidia Blackwell Ultra B300 Chips Discovered Inside China Telecom's State-Owned Network

Thirty-two Nvidia Blackwell Ultra B300 graphics processing units have been located within the network of China Telecom, a state-owned telecommunications giant, despite these specific chips being barred from export to the country by U.S. regulations. The discovery, made by the Israeli security firm Lava, raises fresh questions about how high-end silicon continues to reach Chinese data centers even after strict trade controls have been implemented. These findings confirm that the flow of export-controlled hardware into China remains a persistent challenge for global export enforcement.

The B300 chips are among the most powerful data center GPUs currently available, offering significant performance improvements over their predecessors. Because they are subject to strict export controls, they are theoretically only available to whitelisted buyers who agree not to resell them to sanctioned parties, including those based in China. Yet, the presence of these units in China Telecom’s infrastructure suggests that loopholes or illicit pathways still exist, allowing critical AI hardware to bypass official restrictions.

This incident is not an isolated event. In 2025, reports indicated that Chinese companies managed to smuggle approximately $1 billion worth of Nvidia AI chips into the country in a single quarter. While the B200 model was the primary target at that time, distributors had already begun preparing for the launch of the B300 series. The recent discovery underscores that the proliferation of these advanced chips continues to outpace the effectiveness of current export bans.

Nvidia has responded to the news with characteristic restraint, neither confirming nor denying the specific findings but downplaying their overall impact. In a statement to Tom's Hardware, an Nvidia spokesperson noted that recent court cases have shown that the diversion of controlled U.S. products amounts to less than half of one percent of the total volume. The company described this leakage as "a drop in the bucket" compared to the vast amount of domestic compute resources China already possesses, arguing that with tens of millions of GPUs already in service in approved countries, smuggling remains a non-factor.

The method of detection offers a glimpse into the operational vulnerabilities of modern data centers. Lava identified the chips using Nvidia’s own DCGM Exporter tool, a software utility designed for internal monitoring of server health and performance. Although intended for data center operators, this tool can sometimes be accessed from outside the local network and does not always require password authentication. This technical oversight provided the security firm with the visibility needed to pinpoint the unauthorized hardware.

The motivation for acquiring these chips is clear. Announced in March of last year, the B300 was marketed as a device "built for the age of reasoning," featuring 50% more memory and FP4 compute capabilities than the previous-generation B200. As sophisticated large language models like DeepSeek R1 emerge, the demand for such high-performance silicon has intensified. The B300 was expected to ship by the end of 2025, quickly becoming a highly sought-after asset in the global AI market.

The stakes of these smuggling operations have grown significantly in recent months. In August 2026, nine individuals were indicted in Taiwan for the illegal export of B300 GPUs to China, a move that followed an internal investigation at Supermicro that led to the dismissal of several employees. The smuggling network involved a Taiwanese server trading firm known as Flying Tiget Tech, which held whitelisted status for GPU purchases, highlighting the complexity of the illicit supply chain.

The scarcity of new chips has also driven up the price of older hardware. Nvidia’s A100 servers, which are five years old, have seen their prices triple, with some units commanding up to $82,000. This economic pressure has led some organizations to retool gaming GPUs for AI workloads as a temporary measure. As the debate continues over the efficacy of export controls, the discovery of the B300 chips in China Telecom’s network serves as a reminder that the global race for AI computing power remains fiercely contested.

While it remains unknown exactly how the chips entered the server racks or their precise location within the company’s network, their presence is confirmed. Nvidia is reportedly working on a new tool to help locate its chips in the field, though it has not yet been released. Until such measures are in place, the shadowy market for banned hardware is likely to persist, driven by the high value of these components in the evolving landscape of artificial intelligence.