Core Viewpoint - Nvidia's CFO Colette Kress expressed strong optimism regarding the company's data center business, projecting that revenue from data center chips will exceed the previously forecasted $500 billion by the end of 2026 due to robust demand [1][4][5]. Revenue Projections - Kress indicated that customer interest has been increasing since the $500 billion forecast was made in October 2025, suggesting that the revenue estimate will likely be revised upwards [4]. - The $500 billion target is significantly higher than the market consensus of $4.47 trillion, which is about 12% lower than Nvidia's projection [5]. Product Launch and Performance - Nvidia's new Vera Rubin AI platform has been fully launched, featuring six new chips that significantly enhance inference cost and training efficiency, with initial deliveries expected in the second half of 2026 [6]. - The Rubin GPU boasts a performance of 50 PFLOPS for inference and 35 PFLOPS for training, marking a substantial improvement over the previous Blackwell platform [6][8]. Market Impact - Nvidia's announcements have led to significant stock movements, with storage chip giant SanDisk seeing a nearly 27.6% increase in stock price, while data center cooling system manufacturers experienced declines [9][10][12]. - The comments made by Nvidia's leadership have raised concerns about the long-term positioning of cooling equipment in data centers, particularly with the increasing adoption of liquid cooling technologies [14].
5000亿营收预测太保守!英伟达CFO称“肯定会更高”,黄仁勋称客户需求强劲