The US government has approved Nvidia to sell its advanced AI chips in China, according to the Department of Commerce. This includes the H200, Nvidia’s second-most advanced processor, which was previously under restrictions due to concerns it could bolster Chinese tech and military capabilities.
The Department of Commerce mentioned that shipments are allowed as long as there’s enough supply available in the US. Last month, former President Trump stated that sales could occur to “approved customers” in China, along with a 25% fee on those transactions.
Nvidia expressed support for this decision, highlighting that it could benefit US manufacturing and create jobs. However, there are conditions: Chinese buyers must have proper security measures in place and cannot use the chips for military purposes.
The H200 chip is still a generation behind Nvidia’s most advanced model, the Blackwell processor, which remains banned from sale in China. This ongoing situation reflects the larger geopolitical struggle between the US and China in the AI race. Liu Pengyu, a spokesman for the Chinese embassy, criticized the restrictions, saying they disrupt the stability of supply chains and don’t serve either country’s interests.
Throughout 2025, Nvidia’s CEO Jensen Huang advocated for the sale of these chips, stressing the importance of global market access for America’s competitiveness. Some US officials worry that these chips could end up assisting China’s military, which complicates matters. Despite these concerns, local Chinese firms are likely eager to get their hands on H200 chips, at least until their own technologies improve.
Industry experts point out that while the US may benefit from selling these chips, Nvidia, too, stands to gain financially—even if the profit margins are reduced by the government’s cut. This strategy, proposed by Trump, could influence future negotiations around trade policies in other sectors, according to Marc Einstein from Counterpoint Research.
The global semiconductor landscape has changed dramatically over the years. In the past, the US had a significant edge in chip manufacturing. However, with China’s ongoing push for self-sufficiency in technology, the balance may shift. Current estimates suggest that China aims to produce over 70% of its semiconductor needs domestically by 2025. This could impact both countries, as reliance on foreign technology decreases.
As the situation evolves, keeping an eye on developments in tech policies and international machine learning capabilities will be critical for understanding the future of AI globally.
