U.S. Greenlights Nvidia's Return After 1-Year Hiatus
Nvidia has received U.S. government approval to resume sales of artificial intelligence chips to China, ending an approximate one-year exclusion from the market due to strict export controls. The decision allows the company to export its H200 chip, a powerful GPU developed specifically to comply with American export guidelines. This development marks a significant policy shift after the U.S. tightened restrictions in early 2025, which had previously halted sales and forced Nvidia to take a multi-billion dollar inventory charge.
Following the approval, Nvidia CEO Jensen Huang confirmed at the company's GTC conference that the H200 has been licensed for numerous customers in China and that the company is ramping up manufacturing to meet new orders. This signals Nvidia's formal re-entry into one of the world's largest and most critical technology markets, with sales expected to commence within weeks.
China Reentry Could Unlock $28 Billion Revenue Stream
For investors, the return to China represents a substantial financial opportunity. In fiscal year 2025, before the most recent controls took full effect, sales to China accounted for 13% of Nvidia's total revenue. Based on last year's performance, this could translate to nearly $28 billion in potential annual revenue. CEO Jensen Huang has suggested the long-term opportunity could reach "a couple of 100 billion dollars by the end of the decade."
However, the company faces new challenges. During Nvidia's absence, local Chinese competitors have gained ground, potentially eroding some market share. The pace of revenue recovery will depend on how effectively Nvidia can restart its export logistics and compete with these domestic alternatives. The initial demand from Chinese customers provides a strong basis for optimism that the region will once again become a primary growth driver for the chipmaker.
Tensions Persist as Restricted Chips Surface in China
The approval for H200 sales comes with significant geopolitical complexity. Recent procurement data from March 27 reveals that high-end, restricted Nvidia chips like the A100 continue to be acquired by Chinese entities, including universities with military links. These chips are reportedly being obtained through servers supplied by third-party manufacturers, demonstrating the difficulty of enforcing a complete technological blockade.
This situation has prompted two U.S. senators to call for a pause on all advanced AI chip export licenses to China and its intermediaries. While Nvidia is officially cleared to sell its compliant H200 chips, the ongoing flow of restricted technology highlights the persistent security concerns and regulatory risks shaping the U.S.-China tech rivalry. The policy decision reflects a delicate balance between enabling U.S. commercial leadership and containing China's military AI capabilities.
