Companies are pouring $725 billion into data centers and chips this year.
That is a 77% jump over 2025 and it is starting to break balance sheets.
Free cash flow for the biggest tech giants is plummeting as a result.
Investors are looking at the math and calling it a bubble.
Retail investors are still piling in, but the smart money is exiting.
The valuation gap between hype and reality is at a breaking point.
Opposition to new data centers is sweeping across 49 states.
At least 75 projects worth $130 billion were blocked in Q1 alone.
Communities are pushing back against power and water consumption.
Local resistance is moving into state legislatures with 300 bills filed.
Developers are struggling to find power where the grid is already strained.
Nearly 60% of them are now trying to build their own private power sources.
This is a massive bottleneck for the entire AI expansion timeline.
In 1840s UK “Railway Mania” shares go crazy high, many new lines planned, then big crash. In US 1873 panic start from too much railroad building, overcapacity, banks fail, long depression follow.
Today AI infrastructure (data centers, chips, power) see huge spending. Big tech plan $650-750 billion capex in 2026. Nvidia and others at record highs.
Many articles say it look similar: hype, massive investment, fear of overbuild like railroads. Some worry bubble can burst if returns not come fast.
Library of Congress on 1873 railroad panic: https://guides.loc.gov/business-booms-busts/stock-market
Dame Magazine on AI data center vs railroad bubble: https://www.damemagazine.com/2026/03/17/the-ai-data-center-boom-looks-a-lot-like-the-railroad-bubble/