Nvidia slashes client list in Asia to curb AI chip smuggling

What happened?

Nvidia has implemented a new 'whitelist' of verified companies in Asia, cutting the number of authorized customers for its high-performance GPUs by more than half, according to the Financial Times and reported by Tom's Hardware. This move responds to growing pressure from Washington to prevent its chips from reaching China through third parties. The company has deployed field inspectors to visit data centers, verify contracts, and interview end users to ensure they are legitimate businesses and not shell companies that reship GPUs and servers to China.

This tightening follows a series of police operations in several countries. Notably, the arrest of Supermicro co-founder Wally Liaw, along with two other suspects, for allegedly smuggling $2.5 billion worth of Nvidia hardware to China has been a turning point. According to Tom's Hardware, the accused used sophisticated methods, such as using a hair dryer to move serial numbers between real hardware and thousands of dummy servers, to evade controls. Additionally, a $42 million mansion linked to alleged GPU smugglers was seized in Singapore, and raids were conducted in Taiwan at Supermicro offices and two supply chain partners.

Why is it important?

Since 2022, the US has banned the export of the most advanced AI GPUs to China, but various investigations revealed that Chinese companies still accessed these chips through smuggling via intermediaries in Singapore, Taiwan, and other countries. Nvidia's new whitelist represents a significant strategic shift: the company moves from a reactive to a proactive stance in sanctions compliance. This move not only affects direct customers but also pressures the entire Asian supply chain, including distributors, system integrators, and data centers.

The historical context is key: during the Trump administration, the first restrictions were imposed on Huawei and later expanded to the entire semiconductor sector. With Biden's arrival, sanctions were further tightened, and now under Trump again, the export of H200 GPUs to select customers in China was authorized in December 2025, but Beijing blocked those purchases, betting on domestic semiconductors. However, according to sources cited by the Financial Times, local Chinese production is insufficient: a tech executive stated that all domestic suppliers are sold out and companies are even considering less powerful chips. This underscores China's critical dependence on US technology and the impact of sanctions on its AI development.

Consequences

The tightening of controls has drastically reduced the supply of AI chips in China, making it difficult for local companies to obtain the processors they need. Although President Trump authorized the export of H200 GPUs to select customers in December 2025, Beijing blocked those purchases, leaving Chinese companies with limited options. According to the Financial Times, a tech executive said that all domestic suppliers are sold out and companies are considering even less powerful chips. This could slow down AI progress in China, affecting startups, research centers, and major tech companies like Baidu, Alibaba, and Tencent, which rely on high-performance GPUs to train their models.

For Nvidia, this situation involves reputational and business risk: on one hand, it complies with US regulations, but on the other, it loses a huge market. China accounted for about 20-25% of Nvidia's revenue before sanctions, and although the company has developed chips specifically for the Chinese market (such as the A800 and H800 series, now also restricted), demand remains unmet. This opens opportunities for competitors like AMD, which could gain market share if it manages to navigate sanctions, or for Chinese manufacturers like Huawei with its Ascend chip, though its performance still lags behind Nvidia.

In the global market, the shortage of AI GPUs could delay artificial intelligence projects in Asia, which in turn impacts AI adoption in sectors like healthcare, automotive, and finance. Additionally, regulatory uncertainty could lead companies to diversify their supply sources, investing in local production or open-source alternatives.

What should readers know?

Nvidia has urged its partners to strictly comply with export controls. CEO Jensen Huang stated in May: 'We demand that our partners comply.' This statement reflects a significant shift in Nvidia's compliance strategy, now conducting proactive checks to avoid sanctions. For Asian companies, this means acquiring high-end GPUs will be much harder, and they may be forced to resort to lower-performance alternatives or rely on local Chinese production, which is still insufficient.

It is important to note that the whitelist is not static; Nvidia could adjust it as regulations and investigations evolve. Additionally, companies that want to remain customers will have to undergo rigorous audits, increasing their operational costs. On the other hand, investors should watch how this affects Nvidia's revenue in Asia, as any decline in sales could impact its valuation, currently one of the highest in the tech market.

In summary, Nvidia's new whitelist is a milestone in the technology war between the US and China. While Washington seeks to contain Beijing's technological advance, companies are caught in the middle of sanctions and counter-sanctions. The future of AI in China will depend on its ability to develop competitive semiconductors, but for now, the gap with Nvidia remains huge. Readers should closely follow the next steps from both Nvidia and the governments involved, as the geopolitical landscape continues to shift.