Nvidia has stopped shipping its advanced Hopper series AI chips to China, despite receiving US approval to do so earlier this year. The move, announced by interim CEO Colette Kress during an earnings call, marks a shift from the strategy of former CEO Jensen Huang, who actively lobbied for access to the Chinese market. Nvidia reported $2.4 billion in licensed Hopper chip sales to China in the third quarter, representing about 2% of total revenue, but expects no such sales in future guidance. The decision comes amid heightened scrutiny following reports of Huang’s relationship with a Chinese businesswoman suspected of having ties to Chinese intelligence, and as Nvidia’s board establishes a new panel to oversee national security policies and US export control compliance.
Nvidia’s third-quarter profits reached $65.3 billion, a 105% increase year-over-year, with revenue at $115.8 billion. The company had previously secured permission from the US Department of Commerce in January to sell the H200 chip to China. However, around 600,000 H200 chips were reportedly stuck in Hong Kong earlier this month, as Beijing sought to reduce its reliance on US technology. Kress did not explain the reason for halting shipments. Dilin Wu, a research strategist at Pepperstone, suggested Nvidia is attempting to “draw a clearer line around its China business and reassure regulators that compliance comes first.”
This change in approach contrasts sharply with Jensen Huang’s long-held belief in the importance of the Chinese market. He had previously advocated for maintaining access and sought approval for sales of advanced AI processors, including the Blackwell series, though those requests were denied. Washington continues to ban exports of Nvidia’s most advanced Blackwell and Rubin GPUs to China, citing security concerns related to AI’s potential military applications. The company’s new board-level committee signals that geopolitical considerations will likely play a larger role in future business decisions.
We observe a clear strategic shift at Nvidia. The most likely outcome is a continued reduction of Nvidia’s presence in the Chinese market, prioritizing compliance with US export controls over short-term revenue gains. This will likely be confirmed by further announcements regarding the Blackwell series and a lack of attempts to circumvent restrictions. A less likely scenario involves a negotiated compromise between the US and China, allowing limited sales under stricter conditions, but this would require a significant change in the current geopolitical climate. The stalling of 600,000 H200 chips in Hong Kong suggests Beijing is already preparing for reduced access to US technology, making a return to previous sales volumes improbable. The decision has broader implications for the global AI supply chain. The US is clearly signaling its intent to control access to advanced AI technology, potentially creating a bifurcated market and accelerating China’s efforts to develop its own domestic capabilities. This could lead to increased competition and fragmentation in the AI sector, with implications for innovation and economic growth. Considering the context of Trump’s meetings with Xi and Putin, Nvidia’s move could be interpreted as a deliberate signal of US resolve. While the article does not detail the internal dynamics within the Chinese government, it’s reasonable to assume Beijing views this as a further escalation of tech restrictions and will likely respond with measures to bolster its own semiconductor industry and reduce reliance on US technology.
Read the original coverage
💬 Comments
📜 Comment Policy