AI大跌，背后是黄金坑？还是泡沫？

Core Insights - AI is undergoing a "stress test" as major cloud and chip companies experience significant stock declines despite strong earnings reports [1][2] - Oracle's completion timeline for data centers related to OpenAI has been pushed back from 2027 to 2028, contributing to market concerns [1] - CoreWeave's stock has also fallen, with a notable 46% drop attributed to a major tenant retracting a $150 million investment [1] Financial Performance - Oracle's remaining performance obligations surged 438% year-over-year to $523 billion, driven by demand from tech giants [1] - Broadcom reported Q4 revenue of $18.02 billion, a 28% year-over-year increase, with semiconductor business growth at 34.5% and infrastructure software revenue up 19% [1] - However, both companies reported a negative free cash flow of $10 billion and a cumulative free cash flow of -$13.18 billion over the past 12 months [2] Market Concerns - There are rising warnings about an "AI bubble," with concerns about the sustainability of capital expenditures by tech giants, potential "circular trading" in the industry, and whether future profits can match current high valuations [2] - Howard Marks highlighted that transformative technologies often lead to excessive enthusiasm and investment, resulting in overcapacity and inflated asset prices [2] Economic Perspective - Despite high capital expenditures, major tech firms are seeing an increase in return on invested capital, indicating economic viability [3] - The AI sector is still in its early stages of commercialization, with demand for computing power expected to grow due to advancements in multi-modal models and real-time inference [3] - The current valuation of AI-related companies remains relatively rational compared to historical bubbles, supported by the dual logic of revenue growth and cost reduction [3] Investment Opportunities - The optical module sector is identified as a core support for AI hardware, with increasing industry demand and the rapid scaling of 1.6T products [4] - Domestic computing power sectors are expected to benefit from relaxed export restrictions on advanced computing cards and accelerated IPO processes for local chip companies [4] - Investment strategies focusing on AI-related sectors, such as robotics and computing power, are recommended to capture potential growth opportunities [4]