President Donald Trump announced Monday the United States will allow technology giant Nvidia to export its H200 artificial intelligence chips to "approved customers" in China and other restricted markets, with the U.S. government receiving 25% of the revenue from these sales. The arrangement marks a dramatic policy reversal from the Biden administration's strict export controls and introduces an unprecedented revenue-sharing model in semiconductor trade.
Chinese President Xi Jinping "responded positively" to the proposal, Trump wrote on Truth Social, suggesting the framework emerged from recent discussions between the two leaders. The announcement immediately raises questions about how the deal will be implemented, who qualifies as an "approved customer," and whether the revenue share adequately addresses national security concerns that originally prompted the export restrictions.
Breaking From Biden-Era Export Controls
The Trump administration's decision reverses export restrictions the Biden White House implemented in October 2022 and subsequently tightened multiple times. Those rules specifically targeted advanced AI chips like Nvidia's H100 and H200, preventing their sale to Chinese entities over fears the technology could enhance China's military capabilities, surveillance infrastructure, and AI development.
The H200 represents Nvidia's latest data center GPU, featuring upgraded high-bandwidth memory that delivers substantially faster performance for training and deploying large language models and other AI systems. The chip has been in high demand globally as companies race to build AI infrastructure, with Nvidia struggling to meet orders from U.S. and allied customers.
Nvidia has repeatedly stated that complying with export restrictions cost the company billions in potential Chinese revenue. CEO Jensen Huang previously warned that overly broad restrictions could disadvantage American companies while Chinese competitors develop alternative chips. The Trump deal appears designed to address these concerns while generating government revenue.
The 25% Revenue Share Model
Trump's 25% revenue requirement introduces a novel approach to technology export policy. Rather than outright bans or licensing fees, the government would directly participate in commercial transactions involving strategically important technologies. The mechanism's specifics remain unclear, including whether the 25% applies to gross sales or net profits, how payments would be collected and verified, and what happens if Nvidia or customers fail to comply.
If implemented on gross revenue, the arrangement could generate substantial sums for the U.S. Treasury. Analysts estimate China's AI chip market demand exceeds tens of billions of dollars annually. Even capturing a fraction of that market with a 25% government share would produce significant revenue, though it would also make Nvidia's chips 25% more expensive for Chinese buyers compared to domestic alternatives.
The revenue-sharing model could set precedent for other strategic technology exports, potentially extending to quantum computing components, advanced manufacturing equipment, or biotechnology tools. However, critics argue that revenue considerations shouldn't drive decisions about which capabilities adversaries can access.
National Security Questions Remain
The announcement immediately drew scrutiny from national security experts and congressional lawmakers who supported the original export restrictions. Their primary concern centers on whether financial gains justify providing China access to chips that could accelerate military AI applications, autonomous weapons development, or mass surveillance capabilities.
The Biden administration's export controls explicitly cited these risks, arguing that AI chips enable capabilities threatening U.S. national interests and human rights. The restrictions aimed to maintain America's technological lead in AI by denying competitors access to the most advanced hardware needed for frontier model development.
Trump's "approved customers" language suggests some vetting mechanism will determine which Chinese entities can purchase H200 chips. However, the criteria for approval, oversight of chip usage after sale, and enforcement mechanisms remain unspecified. Previous export control regimes have struggled with enforcement as chips get resold or diverted to restricted entities.
Industry and Geopolitical Implications
For Nvidia, the deal represents a significant victory. The company has watched competitors like AMD and Intel navigate similar restrictions while Chinese firms including Huawei developed domestic alternatives, albeit with performance gaps. Regaining access to Chinese customers, even with the 25% burden, preserves market position and prevents complete ceding of the market to alternatives.
The arrangement signals a broader shift in Trump's approach to U.S.-China technology competition. Rather than decoupling or containment strategies, the administration appears willing to enable controlled engagement when American companies can profit and the government shares revenue. This transactional approach differs fundamentally from the Biden framework's focus on maintaining technological advantages regardless of commercial costs.
For China, purchasing H200 chips—even at effectively 25% markup—may prove preferable to relying entirely on domestic chips that currently lag Nvidia's performance by one or two generations. Access to cutting-edge hardware accelerates AI development timelines and helps Chinese companies remain competitive in global AI markets.
What Happens Next
Implementation details will determine whether the arrangement achieves Trump's stated goals of maintaining American chip leadership while generating government revenue. Key questions include how quickly approved customer lists will be established, whether existing U.S. allies will maintain their own restrictions or follow America's lead, and how Congress will respond to what some members may view as undermining national security for financial gain.
The deal also creates precedent that future administrations must consider. Once advanced chips flow to China through approved channels, reversing access becomes politically and economically more difficult. Whether the 25% revenue share adequately compensates for potential long-term strategic costs of empowering a geopolitical competitor's AI capabilities remains the central question Trump's framework must answer.
