TL;DR:
• Nvidia and AMD will pay US government 15% of China AI chip sales revenue per Financial Times sources
• Deal covers Nvidia's H20 chips and AMD's MI308 processors specifically designed for Chinese market
• Marks dramatic shift from security-based export bans to revenue-sharing licensing model
• Move follows Nvidia's $500B US data center investment pledge that paused chip restrictions
The AI chip war just took an unexpected turn toward pure capitalism. Nvidia and AMD have agreed to hand over 15% of their China AI chip revenues to the US government in exchange for export licenses, according to a Financial Times report citing government sources. The deal transforms national security restrictions into a revenue-sharing arrangement that could reshape how America manages tech exports.
The global semiconductor chess match just became a business transaction. Nvidia and AMD have struck an unprecedented deal with the US government, agreeing to fork over 15% of their revenue from high-end AI chip sales to China in exchange for export licenses, according to Financial Times sources familiar with the arrangement.
The revenue-sharing agreement specifically covers Nvidia's H20 AI chips and AMD's MI308 processors—both engineered specifically for the Chinese market to comply with previous restrictions. Government sources tell the Financial Times that licensing has already begun, marking a dramatic pivot from the Trump administration's initial export ban strategy.
This represents the latest twist in a regulatory whiplash that's sent both companies' China strategies careening between prohibition and permission. The Trump administration initially restricted high-performance AI inference chip sales to China in April, only to pause the ban months later when Nvidia promised Reuters up to $500 billion in US data center investments.
"We follow rules the U.S. government sets for our participation in worldwide markets," an Nvidia spokesperson told TechCrunch. "While we haven't shipped H20 to China for months, we hope export control rules will let America compete in China and worldwide." The carefully worded statement acknowledges the months-long shipping halt while expressing optimism about renewed access.
The policy reversal traces back to broader trade negotiations that extend far beyond semiconductors. Commerce Secretary Howard Lutnick previously linked Nvidia's renewed China access to trade discussions regarding rare-earth elements—critical materials for everything from EV batteries to renewable energy infrastructure that China dominates globally.
For Nvidia, the H20 chip represents a significant engineering compromise. The company specifically designed these processors to meet previous Biden-era restrictions while maintaining enough performance to serve Chinese AI developers. Now, with the revenue-sharing model, Nvidia can monetize that investment while the US government captures a direct stake in the profits.
AMD's inclusion in the deal signals the arrangement extends beyond Nvidia's market dominance. The company's MI308 accelerators, while less prevalent than Nvidia's offerings, represent AMD's push to challenge Nvidia's AI chip supremacy in international markets.
The 15% revenue cut transforms export controls from a binary on-off switch into a graduated taxation system. Rather than completely blocking American companies from the world's second-largest economy, the government now profits directly from their success there. This model could potentially extend to other restricted technologies and countries, creating a new framework for managing strategic competition.
National security hawks aren't buying the compromise. Former government officials and security experts wrote to Lutnick last month urging a complete reversal of chip sale approvals, arguing that revenue sharing doesn't address core security concerns about AI capability transfers.
The timing coincides with intensifying global AI competition as countries race to secure chip access for domestic AI development. China's push for technological self-sufficiency makes American AI chips increasingly valuable bargaining chips in broader trade negotiations.
For investors, the deal reduces regulatory uncertainty while creating a new cost structure for China operations. Both Nvidia and AMD can now plan long-term China strategies around predictable revenue sharing rather than navigating unpredictable export bans.
The revenue-sharing model represents a fundamental shift in how America manages tech competition with China—from prohibition to participation with profit. While critics question whether 15% government cuts address legitimate security concerns, the deal offers both chipmakers and policymakers a middle path between complete restrictions and unfettered access. The real test will be whether this framework can balance commercial interests with national security imperatives as AI capabilities continue advancing.
