英伟达产能大调整：H200让位给下一代芯片Vera Rubin

Core Insights - Nvidia is shifting its manufacturing capacity from the H200 chip to the next-generation Vera Rubin architecture, reflecting the company's latest assessment of the current regulatory environment [1] - The company has paused production of the H200 chip for specific markets, reallocating that capacity to Vera Rubin, which is in high demand from tech giants like OpenAI and Google [1][2] - Nvidia's CFO stated that despite some H200 products receiving export approval, no revenue has been generated from them, indicating a lack of confidence in large-scale sales of the H200 [1] Group 1 - The H200 is an earlier generation AI processor positioned as a compliant product under U.S. export controls, while Vera Rubin is designed for more complex AI systems and is in high demand in the high-end computing market [2] - Nvidia has produced approximately 250,000 H200 chips, and existing inventory is expected to meet demand if only limited orders are approved [2] - A source indicated that Nvidia must pivot to achievable goals rather than waiting in uncertainty, especially given the strong demand for advanced products [2] Group 2 - Nvidia previously sought sales approval for the H200, ramping up production after signals from the Trump administration in December, with expectations of over 1 million orders from customers [3] - The approval process has since stalled, impacting the supply chain that was prepared for H200 deliveries, which were initially planned to start as early as March [3] - Despite the shift in capacity, Nvidia has not completely ruled out the possibility of restarting H200 production if policy changes occur, with existing inventory able to cover demand during a potential three-month ramp-up period [3]