AI-related capital expenditure is poised to rewrite the record books for private-sector infrastructure investment, with projections pointing to staggering sums over the next several years.
Goldman Sachs Research has adjusted widely cited measures of US hyperscaler capital expenditure to produce a more comprehensive global estimate, arriving at $1 trillion of AI-related investment in 2026 alone.
Of that $1 trillion global figure, approximately $581 billion is expected to be deployed within the United States, underscoring how dominant American technology companies remain in driving this buildout.
The consensus of analyst estimates puts US hyperscaler spending at roughly $800 billion this year, a figure that has already become a benchmark in market commentary around the AI investment cycle.
Looking further out, Goldman Sachs baseline aggregate AI capital expenditure estimates suggest approximately $7.6 trillion of capital between 2026 and 2031, spanning compute, data centers, and power infrastructure.
That figure falls within the $4 trillion to $8 trillion range of total capital investment over the next five years that has featured prominently in recent market discussions and analyst reports.
As a share of US GDP, AI-related capital expenditure reached approximately 1.2% in 2025, already exceeding the peak of the telecom buildout that defined the dot-com boom era.
The only comparable private-sector infrastructure investment in American history is the railroad buildout, which peaked at approximately 6% of GDP in the 1880s, setting an almost untouchable historical benchmark.
When adjusted for the shorter useful life of AI chips versus physical infrastructure, current AI spending surpasses even the railroad buildout of the 1860s and 1870s in economic terms.
At a hypothetical $1.25 trillion annual pace, AI investment would represent about 3% of gross domestic product, which would still fall below the peak intensity of the US and UK railroad construction booms of the late 1800s.
A critical structural feature separating this investment cycle from previous booms is that most hyperscaler spending is being self-financed rather than relying on external capital markets.
As of the third quarter of 2025, major technology companies held cash and equivalents totaling approximately $490 billion and generated nearly $400 billion in trailing twelve-month free cash flow.
That financial strength means the AI buildout is not dependent on external financing in the way that the railroad or telecom booms were, reducing systemic financial risk considerably.
The scale and self-funded nature of this investment cycle have prompted analysts to treat it as a genuinely novel phenomenon in the history of American and global capital deployment.
