Beijing has permitted ByteDance and Tencent to each receive approximately 10,000 Nvidia H200 chips at their mainland China facilities in recent weeks, easing restrictions on advanced AI hardware as the country seeks to narrow the gap with U.S. rivals2.
The shipments involved small batches of H200 chips, with Beijing allowing the deliveries to help leading Chinese tech groups in their drive to catch up with U.S. competitors in AI development1. U.S. licenses permit the H200 in China and Hong Kong, according to the Financial Times report.
The move marks a shift in China's posture toward imported AI accelerators. The Financial Times characterized the policy change as China easing limits on Nvidia H200 chips as the AI race escalates.
ANALYSIS The H200 shipments to ByteDance and Tencent represent a constrained but meaningful injection of compute for two of China's largest AI operations — enough to bolster training infrastructure, though the batch sizes remain limited.
That Beijing is facilitating rather than blocking these shipments suggests a pragmatic calculation: enabling domestic AI champions to access frontier-class hardware outweighs whatever leverage restricting imports might provide.
Nvidia Corp. has been navigating a complex geopolitical landscape around chip exports. The company recently disclosed in an SEC filing a $21 billion stake in SpaceX and a $30 billion position in Intel ctx, underscoring the scale of its financial footprint beyond its core GPU business.
ANALYSIS The permitted H200 shipments to ByteDance and Tencent place Nvidia in a position where its hardware continues to flow into China's largest AI operations under existing U.S. licensing frameworks, maintaining a commercial channel that export controls have repeatedly threatened to close.