AI build-out could rival historic rail expansion, CBRE says

The artificial intelligence infrastructure boom is not just another tech spending cycle. According to a landmark new report from CBRE, the world's largest commercial real estate services firm, the current AI build-out could rival the railroad expansion of the 1850s in relative economic scale — and the numbers backing that claim are staggering.

Big tech companies classified as hyperscalers — Google, Amazon, Microsoft, and their peers — are projected to spend $3.7 trillion on AI infrastructure over the next five years. That figure alone would make this the largest private infrastructure investment in modern history. But what makes the comparison to the railroad era particularly apt is how the spending is reshaping the physical landscape of the American economy.

Data center construction today is 12 times what it was in 2020. Manufacturing leasing is up 28% from early 2025, driven largely by industries supporting the tech firms building out AI infrastructure. And U.S.-based AI startups have raised roughly $578 billion in venture capital since 2020, with nearly three-quarters of that flooding in during just the past two years.

The economic footprint is already measurable. CBRE reports that the AI buildout last year was equivalent to nearly half of all U.S. gross domestic product growth, compared with just 8% in 2023 and 2024. That is an extraordinary concentration of economic activity in a single sector, and it raises questions about sustainability that echo the railroad era's boom-and-bust cycles.

Julie Whelan, head of occupier research at CBRE, frames the transformation in historical terms. Like the internet and smartphone revolutions before it, AI is expected to change how jobs work more than how many jobs exist. Employment among entry-level workers in the most AI-exposed occupations has already dipped compared with those in less-exposed roles, but CBRE projects that office-using jobs have historically grown fastest following major technological advancements — and expects the same this time, albeit to a lesser degree.

The commercial real estate implications are immediate and significant. Data centers and industrial properties are the clearest winners, but the ripple effects extend to office demand in major markets with deep pools of skilled talent. Communities hosting data center construction are seeing surges in local economic activity, from construction jobs to increased tax revenue.

Not everyone is optimistic. A Pew Research Center report found that worry outweighs hope when it comes to AI's workforce impact, with roughly a third of workers expecting AI to reduce job opportunities. A Brookings Metro report warned that vital pathways to better-paying jobs for millions of Americans are under threat. And nearly half of college students are reconsidering their majors because of AI uncertainty, according to Gallup and the Lumina Foundation.

Yet 80% of American companies are already using AI in some form, even if 90% report little tangible impact on their business so far. That gap between adoption and impact suggests we are still in the early innings of a transformation that will accelerate as the infrastructure currently being built comes online.

The parallel to the railroad era is instructive in another way. The transcontinental railroad transformed the American economy, but it also produced spectacular failures, speculative bubbles, and a painful period of adjustment before the long-term benefits materialized. The AI build-out is likely to follow a similar pattern — enormous long-term gains, but with significant disruption along the way.

**What This Means For You:** If you work in commercial real estate, data center demand is the clearest growth story of the decade — but the secondary effects on office and industrial markets are where the less obvious opportunities live. If you're a worker concerned about AI displacement, CBRE's data suggests the threat is real for entry-level roles in particular, but that reskilling and AI literacy remain the best defense. If you're an investor, the $3.7 trillion spending forecast means the infrastructure layer — chipmakers, cooling systems, power generation, and construction — may offer better risk-adjusted returns than the AI software companies themselves, which are already priced for perfection. The railroad analogy is worth remembering: the companies that built the tracks often went bust, but the economy that the tracks enabled boomed for decades.