光大证券:予金山云“买入”评级 云服务有望迎来估值重塑
Zhi Tong Cai Jing·2026-01-27 01:52

Core Viewpoint - The strong demand for AI computing power is reshaping the cloud service industry, with a potential end to vicious price competition following AWS's price increase, leading to a dual recovery in profitability and valuation systems for the industry [1] Group 1: Industry Insights - AWS has broken the long-standing trend of only decreasing prices by raising the prices of its AI computing services by approximately 15%, indicating a shift in the pricing strategy within the cloud service sector [2] - Domestic cloud providers are expected to follow suit, potentially stabilizing the profitability of the cloud service industry as AI capital expenditures remain high [2] - The global AI token volume is projected to increase by over 100 times in the next year, leading to upward pressure on hardware costs [2] Group 2: Company Performance - Kingsoft Cloud is expected to achieve a revenue of 2.7 billion yuan in Q4 2025, representing a year-on-year growth of 22% and a quarter-on-quarter growth of 10%, driven by AI-driven public cloud business growth and the traditional peak season for industry cloud [2] - The company is projected to maintain a gross margin of around 15% and an adjusted EBITDA margin of 22%, indicating sustained profitability [2] - Kingsoft Cloud's AI business is anticipated to see a significant revenue increase, with Q4 2025 AI revenue expected to grow by 85% year-on-year to 880 million yuan [4] Group 3: Strategic Positioning - As the only strategic cloud platform within the Xiaomi and Kingsoft ecosystem, Kingsoft Cloud is well-positioned to benefit from the training and inference needs of Xiaomi's self-developed AI models and the integration of numerous terminal devices into AI services [3] - The company has established a neutral cloud vendor position and has covered most leading domestic AI startups, indicating a diversified client base and growth potential in the AI sector [3] - Kingsoft Cloud's AI capabilities are continuously improving, with the recent upgrade of its AI training and inference platform, which is expected to support sustained high growth in AI revenue [4]