博通连续暴跌：一场AI基础设施的再定价！

Core Insights - The article discusses the recent decline in AI-related stocks, particularly focusing on Broadcom's earnings report and its implications for the AI infrastructure business [5][8] - It emphasizes that the decline is not due to weakening AI demand but rather a shift in Broadcom's business model from selling chips to selling complete systems [6][10] Group 1: Broadcom's Earnings Report - Broadcom reported an astonishing demand for AI-related orders exceeding $73 billion over the next 18 months, indicating no signs of slowdown in AI demand [5] - The company is transitioning to a "system sale" model, which means it will be responsible for the entire system's operability rather than just selling individual chips [6] - This shift will lead to a change in profit margins, with the CFO indicating a projected decline in gross margin by approximately 100 basis points in the next quarter due to increased AI revenue and the inclusion of non-Broadcom components in system sales [6][7] Group 2: Market Reactions and Implications - The market is not rejecting AI but is applying a more realistic discount rate to the AI infrastructure business, reflecting the complexities of system delivery and profit distribution [8][9] - The decline in Broadcom's stock has affected a wide range of AI-related companies, as they are viewed as part of the same supply chain and are subject to similar market reassessments [9] - Companies like NVIDIA have not experienced the same decline because they are seen as the endpoint of computing power pricing, benefiting from ongoing demand for AI training and inference [10][11] Group 3: Broader Industry Impact - The article highlights that the importance of components like HBM in AI systems is increasing, and companies like Micron may still see structural profit improvements despite market volatility [11] - Companies in the optical and laser sectors, such as Coherent, are experiencing record orders, but their stock prices may be more volatile as they are perceived as proxies for AI capital expenditure [11][12] - The transition from a focus on growth speed to efficiency, structure, and return paths in the AI sector is becoming crucial as the industry matures [12]