Smuggled Nvidia Chips Found in China

For millions of Americans using smartphones or streaming services, the discovery that high-end Nvidia chips have been smuggled into China does not immediately change the price of a coffee or the speed of their home Wi-Fi. However, it confirms a persistent global issue: the most powerful computer processors used to run artificial intelligence are still finding their way to buyers who are technically restricted from purchasing them. This ongoing flow of hardware suggests that the gap in computing power between the U.S. and China remains narrower than some officials would like to believe.

The revelation comes from an Israeli security firm known as Lava. Researchers at the company identified 32 Nvidia Blackwell Ultra B300 graphics processing units within the network of China Telecom. China Telecom is a state-owned entity responsible for providing telecommunications services across the country. The presence of these specific chips is notable because U.S. export controls have already barred them from being sold to Chinese buyers. Normally, these advanced components are only available to a specific list of approved customers who agree not to resell them to sanctioned parties, including those located in China.

This is not the first time that restrictions on Nvidia’s high-performance hardware have been circumvented. In 2025, reports indicated that Chinese companies had smuggled approximately $1 billion worth of Nvidia AI chips into the country in a span of just three months. At that time, the B200 model was the most sought-after item. However, distributors were already preparing to move the newer B300-series chips as soon as they became available. The current discovery by Lava underscores that the proliferation of export-controlled technology into China continues to be a significant logistical challenge for regulators.

Nvidia has responded to the news with a mix of acknowledgement and downplaying. A spokesperson for the company stated that while governments, researchers, and media outlets have found instances of diverted U.S. products, the volume is "less than one half of one percent" in recent legal cases. Nvidia described this amount as "a drop in the bucket" when compared to the massive amount of domestic computing power China already possesses. The company emphasized that with tens of millions of its GPUs currently in service in approved countries, the smuggling issue remains a minor factor in the broader market.

How the chips arrived in China Telecom’s server racks remains a mystery. Lava has not confirmed the exact location of the processors within the company's vast network, but they were able to detect them 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 often be accessed from outside the data center's secure network. Complicating matters further, the tool does not always require password authentication, creating a potential security gap that allowed the chips to be spotted.

The B300 is highly desirable because it represents a significant leap in capability. Announced in March of last year, this data center GPU offers 50% more memory and FP4 compute power than its predecessor, the B200. Nvidia originally marketed the chip as being "built for the age of reasoning," referencing sophisticated AI models like DeepSeek R1. The chip was expected to ship by the end of 2025, and it quickly became a highly sought-after commodity in the global tech market.

The pressure to acquire these chips has led to notable legal actions in recent months. In August 2026, nine individuals in Taiwan were indicted for the illegal export of B300 GPUs to China. This legal move followed closely on the heels of Supermicro, a major server manufacturer, confirming that it had fired several employees after an internal investigation into smuggling practices. The smuggling operation reportedly involved a Taiwanese server trading firm called Flying Tiget Tech, which held whitelisted status for purchasing GPUs, adding another layer of complexity to the enforcement efforts.

The demand for Nvidia’s silicon has such a profound impact on the market that even older models have seen their prices skyrocket. The company’s A100 servers, which are five years old, have tripled in price and can now cost up to $82,000. This scarcity has driven some buyers to retool gaming GPUs to handle AI workloads where possible, further illustrating the intense competition for available computing power. As Nvidia continues to work on a new tool to help locate its chips, the situation highlights the ongoing tension between global trade controls and the insatiable appetite for AI hardware.

For the average reader, the immediate impact is subtle but meaningful. It suggests that the race for AI dominance is not just about who builds the best software, but also about who can physically secure the best hardware. As the U.S. and China continue to navigate these export controls, the market for high-end GPUs will likely remain volatile. The discovery of these 32 chips is a small data point, but it reinforces the reality that in the world of high-performance computing, borders are not always as effective barriers as they might seem.