Trump allows Nvidia AI chip sales to China for a fee
In a move that has sent ripples through the technology and national security communities, former President Donald Trump has confirmed a unique agreement allowing American chipmakers Nvidia and AMD to resume sales of certain advanced artificial intelligence chips to China in exchange for a 15 percent cut of the revenue paid directly to the U.S. federal government. The announcement, stemming from a recent meeting between Trump and Nvidia CEO Jensen Huang, signals a significant shift in the contentious landscape of U.S.-China tech relations.
President Trump, speaking on the arrangement he dubbed a “little deal,” revealed he initially sought a 20 percent share from Nvidia but ultimately settled on 15 percent during negotiations with Huang. This unusual revenue-sharing model applies specifically to Nvidia’s H20 and AMD’s MI308 chips, which are designed for AI applications. The H20, in particular, was specifically tailored by Nvidia for the Chinese market to comply with previous U.S. export restrictions. Despite its market relevance in China, Trump characterized the H20 chip as “obsolete,” suggesting its sale would not compromise American technological superiority. He further hinted at the possibility of future arrangements for even more advanced chips, such as Nvidia’s cutting-edge Blackwell, provided they are “unenhanced” or “scaled-down” by 30 to 50 percent for the Chinese market, again in exchange for a fee.
The agreement marks a reversal from earlier in the year when the Trump administration had banned H20 chip sales to China in April 2025, only to clear the way for resumed shipments in July, preceding critical trade discussions with Beijing. This fluctuating stance has underscored a perceived incoherence in U.S. chip export control strategy, which historically has been driven by national security considerations rather than economic gain.
The decision has been met with widespread “disbelief” and sharp criticism from national security experts, economists, and bipartisan lawmakers. Critics argue that export controls are a vital tool for safeguarding national security by preventing adversaries from accessing dual-use technologies that could bolster their military capabilities. Monetizing these controls through a revenue-sharing scheme, they contend, sets a dangerous and unprecedented precedent, effectively creating a “pay-to-play” system for access to restricted markets. This approach, many argue, undermines Washington’s global credibility and leverage in economic statecraft, potentially eroding international trust in the consistency of U.S. export policies.
Nvidia, for its part, has maintained that it adheres strictly to the rules set by the U.S. government for its global market participation. The company has expressed hopes that export control policies will continue to allow American firms to compete worldwide, suggesting that continued engagement could keep China reliant on U.S. technology rather than accelerating its indigenous chip development. Both Nvidia and AMD had previously reported significant financial impacts from earlier export restrictions, with AMD facing an $800 million charge and Nvidia anticipating an $8 billion hit.
This latest development echoes previous instances where Trump’s administration sought financial concessions from major corporations in exchange for favorable policy outcomes, such as Apple’s $100 billion investment in U.S. manufacturing to avoid tariffs. However, applying such a model to national security-driven export controls on critical technologies like AI chips presents a novel and contentious challenge to long-standing U.S. foreign policy principles.