Group 1 - The core viewpoint of the article highlights the significant growth in electricity demand driven by advancements in AI technology and the expansion of AI infrastructure, positioning electricity as a key production factor in the digital economy [1] - The issuance of the Invesco Power ETF, which closely tracks the CSI All Share Power Utility Index, aims to provide investors with a convenient tool for exposure to the electricity sector [1] - The CSI All Share Power Utility Index includes 57 stocks from the power utility sector, covering various types of electricity generation such as thermal, hydro, wind, and nuclear power, which helps in diversifying investment risks [1] Group 2 - The electricity sector, as a public utility, is relatively insulated from economic cycles, exhibiting stable profitability and low valuations, with a dividend yield of 2.57% as of December 31, 2025, surpassing the overall A-share market yield of 1.96% [2] - Over the past five years, the CSI All Share Power Utility Index has shown a cumulative increase of 30.30%, indicating robust performance throughout a complete market cycle [2] - The ETF research team at Invesco Power Fund has an average industry experience of over 10 years, with the fund manager, Gong Lili, currently managing various ETF products, showcasing a blend of domestic and international ETF investment management experience [2]

2025年以来，随着AI技术的持续突破，算力爆发式增长推动数据中心等AI基础设施快速扩张，对电力 的需求水涨船高，使其成为数字经济时代的核心生产要素；同时，伴随国内政策对电力等可再生能源的 支持，其在终端能源消费中的占比或将持续提升。在此背景下，电力ETF景顺（代码：159158）的发 行，可谓恰逢其时，该基金紧密跟踪中证全指电力公用事业指数，广泛覆盖多种电力行业，为投资者提 供布局电力的便捷工具。 需求爆发+政策支持，双轮驱动 "算力的尽头是电力"并非短期炒作，而是基于AI算力持续扩张对电力消耗提出巨大需求的长期趋势。一 方面，随着AI大模型不断更新迭代，训练功耗也在持续增加，例如GPT-4.5进行一次训练所消耗的电量 大约是110亿千瓦时，相当于100万个美国家庭一整年的用电量；另一方面，数据中心的扩张对电力的需 求也正快速增长，据IEA预测，到2030年，全球数据中心用电量将增至约945TWh，较2024年 （415TWh）增长一倍以上，期间复合增速约15%。（数据来源：信达证券、天风证券等机构） 然而面对电力需求的增长，海外的电力供应却受限于设备老化、能源转型和制造业回流抢电等因素，以 美国为例，20 ...