Nvidia, AMD to Pay US 15% of China AI Chip Revenue

Chipmakers Nvidia and AMD will remit 15% of their revenue from Chinese AI chip sales to the US government to secure export licenses, an unprecedented arrangement that could unnerve both the American companies and Beijing. This unusual quid pro quo marks a significant development in export control policy, impacting revenue streams and geopolitical dynamics in the technology sector.

Analysis

Nvidia (NVDA) and Advanced Micro Devices (AMD) have entered into an unprecedented arrangement with the US government, agreeing to remit 15% of their revenue from AI chip sales to China in order to secure export licenses. This policy introduces a direct and material cost, functioning as a government-imposed revenue share that will immediately compress gross margins on sales to this critical market. The nature of this "quid pro quo," as described by the Center for Security and Emerging Technology, marks a significant escalation in US export control strategy, moving beyond simple prohibitions to a model that directly captures revenue. This development introduces a new layer of regulatory and geopolitical uncertainty, as it is likely to be met with consternation by both the affected companies and Beijing, potentially risking retaliatory measures and further destabilizing the tech supply chain.

Key Decisions for Investors

• Investors should immediately revise financial models for Nvidia and AMD to account for a 15% reduction in gross revenue from their China AI chip segments, directly impacting earnings forecasts and valuation multiples.
• Monitor for potential retaliatory actions from Beijing, as this unprecedented US policy could provoke countermeasures that further restrict market access or create operational headwinds for US tech firms in China.
• Consider this policy as a potential precedent for other technology sectors with significant China exposure, introducing a new form of regulatory risk that could be applied to other companies seeking export licenses for sensitive technologies.