Guohai Securities·2025-10-21 10:03Investment Rating - The report maintains a "Recommended" rating for the computer industry [1] Core Insights - The cloud computing industry is transitioning towards intelligence, with a significant re-evaluation of industry space driven by AI [5][16] - The global IaaS market is expected to grow significantly, with projections indicating that the domestic IaaS market will approach 800 billion yuan by 2029 [19][21] - Major cloud providers are experiencing high growth in AI-related revenues, which is driving increased capital expenditures [25][51] Summary by Sections Overview of Cloud Computing - The cloud computing industry has evolved through various stages, with the current phase being characterized by the integration of AI, particularly following the rise of large models like ChatGPT in 2023 [16][19] - The global top four cloud providers are seeing rapid growth in AI business revenues, with orders continuing to rise [5][51] IaaS Market Dynamics - The IaaS model provides essential IT resources such as computing, storage, and networking, allowing users to pay based on actual resource usage [10][12] - The domestic IaaS market is projected to reach 4,355.2 billion yuan in 2025, with a year-on-year growth of 17.1% [19][21] Competitive Landscape - The global IaaS market remains highly concentrated, with the top four players (Amazon, Microsoft, Google, and Alibaba) holding a significant market share [22][24] - The competition among these giants is intensifying as they invest heavily in AI infrastructure and capabilities [24][25] Growth Drivers - The demand for AI IaaS is reshaping the cloud computing landscape, with a shift from traditional resource scaling to intelligent capabilities [48][56] - The focus is shifting towards application layers such as AI PaaS and MaaS, driven by the increasing need for AI applications [66] Future Outlook - The report anticipates that the inference market will become a new growth driver, with commercial opportunities moving towards application layers [57][65] - The evolution of large models towards enhanced reasoning and multi-modal capabilities is expected to unlock new application demands [60][61]