四巨头“烧钱凶猛”，非美和二线云厂被低估，GB200良率提升！大摩对AI服务器非常乐观

Group 1 - A global cloud infrastructure competition driven by AI is rapidly intensifying, with significant capital expenditure increases forecasted for major cloud service providers [1][2] - Morgan Stanley predicts that the combined capital expenditure of Amazon, Google, Meta, and Microsoft will reach $359 billion in 2025, a 57% year-over-year increase, and $454 billion in 2026, a 26% increase [2][5] - The report highlights that the demand from non-US regions and Tier 2 cloud service providers is likely underestimated, indicating a broader market potential [1][7] Group 2 - The capital expenditure growth for the four major US cloud providers is expected to accelerate, with a projected $100 billion in Q4 2025, a 39% year-over-year increase [2] - The total capital expenditure for the top 11 global cloud service providers is expected to reach $445 billion in 2025, significantly higher than previous forecasts [2] Group 3 - By 2026, capital expenditure as a percentage of revenue for these companies is expected to exceed 20%, setting a new historical high [5] - The demand for AI servers, particularly from Tier 2 cloud service providers, is anticipated to grow significantly, with potential capital expenditure increases starting in the second half of 2026 [7] Group 4 - Supply chain issues are easing, with improvements in the assembly yield of Nvidia's GB200 chips, which supports the growing demand for AI servers [8] - Major projects like "Stargate," involving OpenAI, SoftBank, and Oracle, are moving from planning to execution, indicating a shift from order-based to project-based demand [8]