“中国可能都做不到”！AI泡沫充斥得州：超220GW大项目申请到2030年入电网

Core Viewpoint - Texas is experiencing a massive data center construction boom driven by AI, raising concerns about a potential bubble due to the overwhelming demand that cannot be met by existing infrastructure [1][3]. Group 1: Data Center Demand and Capacity - ERCOT reported that large projects applying to connect to the Texas grid before 2030 exceed 220 GW, with over 70% being data center projects [1]. - The 220 GW demand is more than double Texas's record summer peak demand of approximately 85 GW and far exceeds the state's quarterly total generation capacity of about 103 GW [1]. - In 2023, the number of large projects applying for power access nearly quadrupled, with over half of the projects representing about 128 GW of potential demand growth not yet submitted for ERCOT review [4][5]. Group 2: Infrastructure and Feasibility Concerns - Experts express skepticism about the feasibility of building sufficient infrastructure to meet the projected demand, with Joshua Rhodes stating that the numbers appear absurd and likening the situation to a bubble [2][3]. - Actual projects integrated into the grid or approved by ERCOT amount to only about 7.5 GW, which is still significant, equivalent to nearly eight large nuclear power plants [5]. - Texas has implemented measures to distinguish serious data center projects from speculative ones, including requiring developers to pay for preliminary studies and provide proof of site control [6]. Group 3: Financial and Credit Market Implications - The data center construction boom has raised concerns about potential oversupply, with some centers possibly becoming unprofitable, leading to bankruptcies [3][4]. - The credit market is witnessing a frenzy, with significant debt financing for data center projects, raising fears of a bubble that could impact equity and credit participants [8][9]. - Reports indicate that at least $175 billion in credit transactions related to data centers have been completed in the U.S. this year, with some borrowers seeking loans exceeding project costs [9].