Nvidia’s Banned B300 Chips Discovered in China Telecom’s Infrastructure

The global race for advanced computing power has long been a battleground for geopolitical strategy, but a recent discovery suggests the lines between sanctioned and unsanctioned hardware are blurring. Israeli security firm Lava has announced the identification of 32 Nvidia Blackwell Ultra B300 graphics processing units within the network of China Telecom, a state-owned telecommunications provider operating within China. The presence of these specific chips is significant because U.S. export controls have explicitly barred their sale to Chinese entities, raising immediate questions about how they bypassed regulatory filters to reach a major domestic data center.

The core of the controversy lies in the mechanism of sale. Nvidia’s high-performance GPUs are typically sold only to whitelisted buyers who contractually agree not to export the hardware to sanctioned parties, including those in China. Despite these strict protocols, the B300 chips—part of the Blackwell Ultra series—have surfaced in a Chinese data center, indicating a potential gap in the oversight process. This finding aligns with a broader trend where Chinese firms have persistently sought to acquire Nvidia’s top-tier silicon, a challenge that has persisted despite repeated attempts by U.S. policymakers to tighten export restrictions.

The discovery is not an isolated incident in the history of chip smuggling. In 2025, reports indicated that Chinese companies managed to smuggle approximately $1 billion worth of Nvidia AI chips into the country over the span of just three months. At that time, the B200 series was the primary target of distributors, but many were already positioning themselves to sell the superior B300-series as soon as it became available. The recurrence of these findings underscores the difficulty in containing the flow of advanced technology, even when formal bans are in place.

Nvidia has responded to the report with a mixture of acknowledgment and minimization. While the company did not deny Lava’s findings, a spokesperson characterized the diversion of controlled products as a "drop in the bucket" compared to the vast amount of domestic compute capacity China already possesses. The spokesperson noted that in recent court cases, the diversion of U.S. products was found to be less than half of one percent, suggesting that with tens of millions of GPUs already in service in approved countries, the impact of this specific batch is negligible. The company maintains that smuggling remains a non-factor, largely due to enforcement efforts by governments on both sides of the Pacific.

However, the specifics of how these chips were located remain somewhat opaque. Lava identified the 32 B300 units using Nvidia’s own DCGM Exporter tool, a software utility installed by data center operators to monitor server health and performance. Although designed for internal use, this tool can often be accessed from outside the data center’s network and frequently lacks mandatory password authentication, creating a potential security loophole. It is currently unknown exactly where within China Telecom’s infrastructure the chips are housed, nor can Lava confirm the precise location within the company’s broader network.

The desire for the B300 is driven by its substantial performance advantages over its predecessor. Announced in March of the previous year, the chip offers 50% more memory and FP4 compute capability than the B200. Nvidia marketed the B300 as "built for the age of reasoning," highlighting its suitability for sophisticated large language models like DeepSeek R1. Given these capabilities, the chip became a highly sought-after asset, with shipments expected by the end of 2025, making it a prime target for those looking to bolster their AI infrastructure without waiting for official import approvals.

The smuggling of these advanced chips has led to tangible consequences for the supply chain. In August 2026, nine individuals in Taiwan were indicted for the illegal export of B300 GPUs to China. This legal action followed closely after Supermicro, a major server manufacturer, confirmed it had fired several employees following an internal investigation into smuggling activities. The illicit trade involved a Taiwanese server trading firm named Flying Tiget Tech, which held whitelisted status for GPU purchases, illustrating how legitimate channels can be exploited to move hardware across borders.

The high demand for Nvidia’s silicon has also distorted the secondary market. Older models, such as the five-year-old A100 servers, have seen their prices triple, with units now commanding up to $82,000. This scarcity has forced some organizations to retool gaming GPUs to handle AI workloads where possible. While Nvidia insists that the current findings do not significantly alter the landscape of AI compute, the continued presence of banned chips in major Chinese networks suggests that the enforcement of export controls remains a complex and ongoing challenge for the industry.

As the technology sector continues to navigate these geopolitical currents, the discovery of the B300 chips in China Telecom’s network serves as a reminder of the persistent gaps in global trade regulations. For observers in the U.S. and South Central states, the story highlights the critical role that advanced hardware plays in the AI arms race, and the lengths to which various actors will go to secure a competitive edge. Whether this incident will lead to stricter enforcement or further adjustments in export policy remains to be seen.