Core Viewpoint - The construction of AI computing power centers is on the rise, and the demand for solid-state transformers (SST) is expected to grow due to their efficiency, compact size, flexibility, and prefabricated design, making them a preferred solution for future data center power supply architectures [1][2]. Group 1: AI Power Supply Architecture - The upgrade of AI power supply architecture to 800V HVDC is anticipated to make SST the optimal technical route in the long term [2]. - Major cloud service providers have raised capital expenditure expectations, accelerating the construction of AI data centers (AIDC) [2]. Group 2: Performance Advantages of SST - SST can improve overall efficiency by over 3 percentage points compared to traditional UPS power supply chains, reduce footprint by over 60%, and shorten installation and debugging cycles by 75% [3]. - SST is adaptable to green electricity integration and has strong advantages in lifecycle costs [3]. Group 3: Cost Reduction Potential - Although the initial investment cost for SST is currently high, there is significant potential for overall cost reduction as core components such as power semiconductor devices and amorphous nanocrystalline materials mature [3]. Group 4: Industry Landscape - Global leaders in AIDC, such as Eaton and Delta, have early reserves in SST solutions, while domestic companies are entering the market from traditional power equipment, AI power supply, and solar storage charging sectors [4]. - Some domestic companies have already applied SST products in projects involving AC/DC hybrid microgrids, with China West Electric having products used in data center projects [4]. Group 5: Recommended Stocks - Recommended stocks include Sifang Co., Ltd. (601126.SH), China West Electric (601179.SH), Jinpan Technology (688676.SH), and other related companies in the industry chain [5].
中金:AI供电架构向800V HVDC升级 长期看SST有望成为最优技术路线