证券研究报告 | 行业周报 gszqdatemark 2025 08 03 年 月 日 电子 周观点：从巨头财报周看最新 Capex CSP 相继上调资本开支指引，高度重视 AI 产业链变革及业绩爆发机遇。 谷歌、Meta、亚马逊相继上调全年资本开支预期，谷歌上调 25 年 100 亿美元资本开支至 850 亿美元，Meta 上调 25 年资本开支预期至 660- 720 亿美元，亚马逊表示目前需求已经超过了供给，上调 25 年资本开 支至 1200 亿美元左右，商业模式角度，微软 Azure 连续大幅超预期， 未来资本开支更多投向短周期资产，我们认为主要系看到了 AI 大范围 的需求以及商务模式正逐步跑通。我们认为目前 AI 大时代已至，大趋 势下，把握行业大 beta，高度重视 AI 产业链变革及业绩爆发机遇。 具体来看： 1）Meta 再次上调资本开支预期至 660-720 亿美元，2026 年预计 Capex 继续大幅增长：本季度资本开支 170 亿美元，2025 年 Capex 指引 660-720 亿美元（前值 640-720 亿），连续上调资本开支，2026 年预计 Capex 继续大幅增长，以 ...

Investment Rating - The report maintains an "Overweight" rating for the electronic industry, indicating a positive outlook for the sector in the next 6 months [1]. Core Insights - North American major CSP manufacturers have optimistic AI investment trends, with significant capital expenditures planned for AI development. Meta's Q1 capital expenditure reached $13.96 billion, a 103.7% year-on-year increase, while Google and Amazon also reported substantial increases in their capital expenditures [1][2]. - AI-related businesses and traditional businesses empowered by AI are rapidly growing, driving demand for inference computing power into a high-growth cycle. Microsoft's Azure cloud service revenue grew by 33% year-on-year, with AI contributing 16% to this growth [2][3]. - The demand for cost-effective computing power is accelerating the deployment of self-developed ASIC chips by CSPs. Google, Amazon, and Microsoft have all introduced new AI chips that significantly reduce inference computing costs, indicating a shift towards ASIC technology in the industry [3]. Summary by Sections Section 1: AI Investment Trends - Major CSPs are increasing their capital expenditures for AI, with Meta, Google, Microsoft, and Amazon all reporting significant year-on-year growth in their investments [1][2]. Section 2: Growth in AI-Driven Services - The rapid growth of AI-related services is evident, with Microsoft's Azure cloud service and Google's AI-driven advertising showing substantial revenue increases [2][3]. Section 3: ASIC Chip Development - CSPs are focusing on developing ASIC chips to lower inference computing costs, with notable advancements from Google, Amazon, and Microsoft in their respective AI chip technologies [3].