China’s Rejection of Nvidia H200 Chips and the Implications
China has reportedly decided against purchasing Nvidia Corp.’s H200 AI chips, a move that White House AI czar David Sacks attributes to the country’s desire for semiconductor independence. This decision highlights a growing trend in China’s strategy to develop its own semiconductors rather than rely on foreign technology.
Sacks made these comments during an interview, citing an unspecified news article he had read. He suggested that China is not interested in the H200 chips because it wants to achieve self-sufficiency in semiconductor production. This aligns with broader efforts by the Chinese government to reduce its dependence on foreign chipmakers like Nvidia.
In a social media post, Sacks referenced a Financial Times report indicating that China might impose restrictions on the use of H200 chips through a local approval process. This would require Chinese buyers to justify their purchases, potentially limiting access to the technology.
The implications of this decision are significant for Nvidia. The company had previously removed China from its forecasts, but CEO Jensen Huang estimated that the Chinese market could be worth $50 billion this year. Analysts suggest that if China accepts the H200 chips, they could generate up to $10 billion in annual revenue.
Nvidia has stated that it continues to work with the administration on obtaining licenses for the H200 chips for vetted customers. However, the company has not yet reported any results from these efforts. A spokesperson noted that three years of overbroad export controls have benefited American competitors and cost taxpayers billions of dollars.
China’s stance on semiconductor independence is further emphasized by a reported package of incentives worth as much as $70 billion aimed at supporting its local chipmaking industry. This move underscores Beijing’s commitment to reducing reliance on foreign chipmakers and bolstering domestic companies like Huawei and Cambricon Technologies Corp.
The decision to allow the H200 chips to be exported to China was based on an assessment that Huawei, Nvidia’s primary competitor in China, offers AI systems with comparable performance. For instance, Huawei’s Cloud Matrix 384 platform links hundreds of processors together to compensate for lower individual chip performance. Some U.S. officials viewed the H200 as a compromise from Nvidia’s earlier push to export a version of the Blackwell chip to China.
Earlier this year, China rejected the H20, a less capable chip that Trump had allowed for export. Despite the recent policy change, Beijing has not publicly agreed to allow imports of the H200 products. There has also been no official rejection, leaving the situation uncertain.
Nvidia CEO Jensen Huang expressed uncertainty about whether China would accept the H200 chips. On Monday, President Donald Trump mentioned in a Truth Social post that President Xi Jinping of China responded positively to the possibility of H200 export approvals.
This ongoing development highlights the complex dynamics between the U.S. and China in the realm of technology and trade. As both nations navigate their strategic interests, the future of AI chip exports remains a critical point of contention.