America's artificial-intelligence investment boom has grown so enormous that one of the most revealing comparisons now comes from an unlikely place: Britain's railway frenzy of the 1840s.
The spending wave across Big Tech has become so large that, measured against the size of the economy, only the UK railway mania of the 1840s was bigger, according to Ameriprise Financial.
The Spending Wave Stretches Across Big Tech
Amazon.com Inc (NASDAQ:AMZN), Microsoft Corp (NASDAQ:MSFT), Meta Platforms Inc (NASDAQ:META), Alphabet Inc (NASDAQ:GOOGL)(NASDAQ:GOOG) and Oracle Corp (NYSE:ORCL) are expected to invest a total of $729 billion on AI in 2026. In 2027, Ameriprise Financial expects that figure to hover around $1.1 trillion.
Put against the size of the economy, the numbers become even more striking.
AI-related capital expenditure accounted for more than one-quarter of all U.S. private fixed investment in the second quarter, Ameriprise said, citing Oxford Economics. Cumulative AI capex has already added roughly 2% of U.S. GDP over the past three years, while hyperscaler investment alone is projected to reach 2.5% of GDP next year.
That is where the railways come in.
The Railway Comparison Is Unusual
Looking across five major technology-driven investment booms spanning roughly two centuries, Ameriprise highlighted an extraordinary comparison: "only the UK railway mania of the 1840s was larger by this comparison."
The comparison is not simply about how much money is being spent. It raises a more uncomfortable question about what happens when investment in transformative technology races ahead of the returns generated by that investment.
Ameriprise noted that technological booms have historically tended to peak after three to five years and can end abruptly. The second UK railway bust cut UK GDP by around 6% relative to trend by 1850, while U.S. railway company defaults surged in the 1870s following another wave of heavy investment.
AI Needs to Show the Math
One important lesson is that transformative technologies can change the economy while the companies financing the buildout still struggle to earn enough on the enormous sums they spend.
That question is becoming more important as AI investment continues to accelerate. Ameriprise said hyperscaler investment this year is approaching 50% of revenue, while the group's combined free cash flow is estimated to fall from $200 billion in 2025 to negative $40 billion this year.
Oxford Economics estimates that generating a 15% to 20% return on roughly $3.8 trillion of cumulative AI investment from 2024 through 2028 would require approximately $570 billion to $800 billion in additional annual profit.
For investors, that makes the next phase of the AI trade less about whether companies will keep spending and more about whether that spending starts producing measurable revenue, profits, free cash flow and productivity gains. With the investment wave still accelerating, the gap between AI's promise and the returns needed to justify the buildout is becoming a number investors cannot easily ignore.