为了人工智能（AI）的发展和普及，构成支撑的基础设施投资正趋于过热。从过去的历史中我们可以 看到，改变社会的大型基础设施建设都是在经历多次失败的情况下完成的。开始有人将最近的AI热潮 与投资热之后出现众多失败者的19世纪"铁路热潮时代"联系起来。 "搁浅资产化"的风险 "就像其他技术基础设施的投资周期所发生的那样，一些投资者或将受到严重伤害"，这是正处于AI投资 漩涡之中的OpenAI首席执行官（CEO）萨姆·奥尔特曼的话。 如果随着技术创新的推进，基础设施建设的成本降低，"任何人都将能够在笔记本电脑上使用个人的通 用AI（AGI）"（奥尔特曼）。这意味着过剩的基础设施投资将成为无法产生预期收益的搁浅资产。 开发AI模型的OpenAI、谷歌、Meta和亚马逊等"hyperscaler（超大规模云服务商）"正在进行前所未有的 大规模投资。美国麦肯锡估算，到2030年将向数据中心等投入近7万亿美元的巨资。 "可能难以置信，但互联网普及时期的投资并没有那么沉重"，OpenAI的首席财务官（CFO）莎拉·弗莱 尔（Sarah Friar）10月下旬在被称为"沙漠达沃斯"的沙特阿拉伯国际金融会议上登台，认为目前的状 况 ...

Core Viewpoint - The article discusses the potential risks associated with the current AI investment boom, drawing parallels to historical investment frenzies, particularly the 19th-century railway boom, which led to numerous failures despite initial enthusiasm [2][7][10]. Group 1: AI Investment Landscape - The infrastructure investments supporting AI development are becoming overheated, reminiscent of past investment bubbles that resulted in significant failures [2][7]. - OpenAI's CEO, Sam Altman, warns that some investors may suffer severe losses due to the risk of "stranded assets" as technology advances and infrastructure costs decrease [3][5]. - Major companies like OpenAI, Google, Meta, and Amazon are making unprecedented investments, with McKinsey estimating nearly $7 trillion will be invested in data centers by 2030 [7]. Group 2: Historical Context and Comparisons - The article compares the current AI investment climate to the railway investment boom, highlighting that massive capital was required for infrastructure, leading to many bankruptcies during the 19th century [7][8]. - In the U.S., 55 companies went bankrupt in 1877 alone, with further failures in subsequent years due to overinvestment and deteriorating returns [8]. - The financial turmoil during the railway boom led to advancements in financial technology and the establishment of modern investment banking practices [8]. Group 3: Financial Viability and Future Outlook - The reliance on developed countries' finances to support massive AI infrastructure investments is questioned, as current funding sources are insufficient [11]. - Meta announced a capital expenditure of $72 billion for the year, with projections for further increases, leading to a significant stock sell-off [11]. - Research indicates that approximately 95% of organizations are currently unable to profit from generative AI investments, raising concerns about the sustainability of the current investment trend [11].