摩根大通最新报告《AI资本开支与二季度前瞻》 JPMorgan's Latest Report: "AI Capex and Q2 Earnings Preview"
市场前瞻 Market Outlook
摩根大通最新报告《AI资本开支与二季度前瞻》核心观点：AI投资周期还没有走到尽头，但市场的关注焦点已经悄然转向——从"比谁烧钱更多"变成"比谁能把钱变成回报"。本轮财报季最大的看点，不再是资本开支有没有继续增加，而是这些AI投入究竟能不能实实在在转化为收入和利润的增长。
JPMorgan's latest report, "AI Capex and Q2 Earnings Preview," argues that the AI investment cycle is far from over, but the market's focus has quietly shifted — from "who is spending the most" to "who can turn that spending into returns." The biggest question this earnings season is no longer whether capital expenditure keeps rising, but whether AI investment can genuinely translate into revenue and profit growth.
一、核心判断：AI资本开支持续扩张，但市场关注重心已转移
I. Core View: AI Capex Keeps Expanding, but Market Focus Has Shifted
摩根大通认为，微软、谷歌、Meta、亚马逊四大云计算巨头，本轮财报大概率不会明显下调AI预算，资本开支仍将维持高位。
JPMorgan believes the four hyperscalers — Microsoft, Google, Meta, and Amazon — are unlikely to meaningfully cut AI budgets this earnings season, with capex expected to remain at elevated levels.
但市场的注意力已经从单纯盯着资本开支规模，转向更实际的三个问题：AI是否真正拉动了收入增长、AI投入的回报率究竟如何、云业务与AI产品的商业化进展到了哪一步。因此，管理层这次在财报电话会上如何表态未来AI投资回报，重要性已经超过资本开支这个数字本身。
However, market attention has moved beyond simply tracking the size of capex, toward three more practical questions: whether AI is genuinely driving revenue growth, what the actual return on AI investment looks like, and how far cloud and AI products have progressed toward commercialization. As a result, how management frames future AI investment returns on this quarter's earnings calls now matters more than the capex figure itself.
二、二季度业绩预期依然强劲
II. Q2 Earnings Expected to Remain Strong
摩根大通预计，科技板块二季度盈利增速约64%，非科技板块约37%，标普500整体盈利增速预计在12%左右。支撑这轮盈利增长的核心动力仍然是云计算业务扩张、AI相关需求旺盛，以及广告业务保持韧性。
JPMorgan forecasts Q2 earnings growth of roughly 64% for the tech sector, about 37% for non-tech sectors, and around 12% for the S&P 500 overall. The core drivers behind this growth remain cloud business expansion, strong AI-related demand, and resilient advertising revenue.
三、AI资本开支短期仍有支撑，主要来自三方面
III. Three Factors Continuing to Support AI Capex in the Near Term
第一，超大型云厂资本开支依旧维持高位：预计2026年，谷歌资本开支约3000亿美元，亚马逊约3000亿美元，Meta约2000亿美元，整体AI投资强度不减。
First, hyperscaler capex remains at elevated levels: for 2026, Google's capex is projected at roughly $300 billion, Amazon's at roughly $300 billion, and Meta's at roughly $200 billion, with overall AI investment intensity holding steady.
第二，主权AI成为新增需求来源：除了美国几大云厂之外，中东、欧洲、亚洲多国政府正在推动主权AI建设，成为新一批GPU需求来源，意味着AI需求正在从四大云厂向外扩散。
Second, sovereign AI has emerged as a new source of demand: beyond the major U.S. cloud providers, governments across the Middle East, Europe, and Asia are pushing forward sovereign AI initiatives, becoming a fresh source of GPU demand — meaning AI demand is now spreading well beyond the four hyperscalers.
第三，AI推理需求快速攀升：随着Agent、Copilot以及各类企业AI应用普及，推理端需求正快速增长，为GPU需求提供进一步支撑。
Third, AI inference demand is rising quickly: as agents, Copilot-style tools, and enterprise AI applications become more widespread, inference-side demand is climbing rapidly, providing further support for GPU demand.
四、增速放缓不等于周期结束
IV. Slower Growth Does Not Mean the Cycle Is Over
摩根大通认为，未来几年AI资本开支的增速可能逐步回落，但增速放缓不等于投资停止，更可能是从高速扩张转入更理性的增长阶段，AI产业链不会因为增速放缓就立刻陷入衰退。
JPMorgan believes AI capex growth rates may gradually moderate over the next few years, but slower growth does not mean spending stops — it more likely marks a shift from rapid expansion toward a more measured growth phase. The AI supply chain is unlikely to fall into decline simply because growth rates ease.
