Core Viewpoint - The report from CICC highlights that as the capabilities of large models continue to enhance, the demand for AI inference computing power is expected to surge, driving growth in AI hardware demand [1] Group 1: AI Investment Directions - The report suggests focusing on three main investment directions in AI: 1) New hardware technologies, with an expected increase in the penetration of cost-effective and energy-efficient AI ASICs, driven by high-performance and low-power demands [2] 2) Domestic production, as the resumption of H20 deliveries and the continuous improvement of the domestic cloud AI chip supply chain are expected to support domestic AI server demand [2] 3) AI applications, where innovations at the terminal level are accelerating, leading to growth in IoT device connectivity and expansion in the communication module market [2] Group 2: Telecommunications Capital Expenditure - Telecom capital expenditures are increasingly leaning towards computing power networks, with a projected 9.1% year-on-year decrease in capital expenditure by the three major operators in 2025, totaling 289.8 billion yuan [2] - The report recommends focusing on opportunities in 5G-A deployment, accelerated 6G technology research, and the growing demand for new types of optical fibers driven by AI in data centers [2] Group 3: Profit Forecast and Valuation - The report maintains profit forecasts, ratings, and target prices for several companies, including Zhongji Xuchuang, Broadcom, Guangxun Technology, and others [3]
中金：AI商业化加速 关注算力主线和政策性机遇
智通财经网·2025-08-13 08:13