摩根士丹利报告：AI 数据中心初创企业 2026 年融资达 335 亿美元，八成投向电力与算力效率
据摩根士丹利最新报告统计，截至 2026 年 10 月初，美国专注于解决 AI 数据中心电力与算力瓶颈的初创企业今年以来已累计融资 335 亿美元，接近 2025 年全年的两倍。报告指出，其中超过八成的资金投向了电力供应保障与计算效率提升两个方向，显示资本关注的重点正在从「建多少机房」转向「电从哪里来、算力如何更省电」。
报告认为，这批初创企业面临的考验已经发生变化：能否真正把规划中的产能落地投产，比融到多少钱更能决定成败。随着大型科技公司持续扩建 AI 基础设施，电网接入、冷却与能源管理成为产业链上最紧缺的环节，也成为创业公司最集中的赛道。
这一趋势与今年以来美国多地围绕数据中心选址、取水与电价的公开讨论相互印证：从地方政府的暂停令到公用事业公司的供电协议，电力已经成为 AI 扩张最直接的约束条件。摩根士丹利在报告中提醒，融资规模创新高的同时，项目交付能力将成为下一阶段分化的关键。对华人科技从业者与投资者而言，这批围绕电力与效率的初创公司，正是观察美国 AI 产业链下一轮机会与风险的窗口。
图片来源：数据中心建筑群实景，CC BY-SA 3.0（Wikimedia Commons，摄影：123net）
AI Data Center Startups Raised $33.5 Billion in 2026, Morgan Stanley Report Finds
U.S. startups working on power and computing constraints for AI data centers have raised $33.5 billion so far in 2026, nearly double the total for all of 2025, according to a Morgan Stanley report covered in early October. More than 80 percent of the funding targeted power availability and computing efficiency, the report found, suggesting investor attention is shifting from how many facilities get announced to whether they can actually be powered and run efficiently.
The report said commercial viability now depends on capacity actually deployed, not just capital raised. As large technology companies keep expanding AI infrastructure, grid access, cooling and energy management have become the scarcest links in the supply chain — and the most crowded field for startups.
The findings line up with a year of public debate across the United States over data center siting, water use and electricity prices, from local moratoriums to utility supply deals. Morgan Stanley cautioned that with fundraising at record levels, delivery capacity will be the factor that separates winners from the rest of the field.
Photo: 数据中心建筑群实景，CC BY-SA 3.0（Wikimedia Commons，摄影：123net）
图片来源：数据中心建筑群实景，CC BY-SA 3.0（Wikimedia Commons，摄影：123net）
