英伟达下一代Rubin芯片已流片,未来5年数据中心资本支出超3万亿

Core Insights - Nvidia is transitioning to a new computing platform that will be relevant for the next few decades, moving away from a standard computing platform that has existed for 20-30 years [1][4] - The company is preparing to launch the Rubin architecture, which will consist of six chips that have already been taped out [1] - Nvidia's CFO, Colette Kress, highlighted significant demand for computing power, particularly driven by inference models, indicating a robust market for their products [3] Group 1: Financial Performance and Projections - In Q2 of fiscal year 2026, Nvidia reported a GAAP gross margin of 72.4%, with an outlook for Q3 gross margin at 73.3% [5] - Nvidia anticipates capital expenditures related to data center infrastructure to reach $3 trillion to $4 trillion over the next five years, emphasizing the importance of this market [4] - The company has not observed significant changes in the lifespan and replacement cycles of deployed chips, with many clients maintaining older Hopper architecture chips due to their high performance [4] Group 2: Market Demand and Supply Chain - Nvidia has seen thousands of megawatts of demand related to the upcoming Rubin architecture before its market entry [3] - The successful operation of the Blackwell Ultra version allows Nvidia to focus on improving gross margins through product mix and cost management [5] - Nvidia's supply chain has been strengthened by long-term relationships with suppliers, which have adapted to meet the company's evolving production needs [5]