i
News
News · 2026-10-11

Data-center construction jumps 64.5% as other sectors retreat

@neuronium_ai @neuronium_ai

A new chart based on data from Harvard Kennedy School economist Jason Furman shows data-center construction surging this year while spending across nearly every other major construction category falls. The gap is striking: data centers are up 64.5%, while manufacturing construction is down nearly 24% compared with 2025. The boom is increasingly concentrated in AI infrastructure, even as investors question whether current levels of AI spending can last.

Cover: Data-center construction jumps 64.5% as other sectors retreat

A boom with few beneficiaries

64.5%data centers
4.1%energy
−5.6%multifamily housing
−10.1%health and retail
−12.6%education
nearly −24%manufacturing

Multifamily construction fell 5.6% over the past year. Spending on health facilities and commercial retail dropped 10.1%, while educational buildings declined 12.6%. Energy was the lone other category in positive territory, up 4.1%.

Manufacturing is a particularly awkward contrast: it has been described as a target for the Trump administration, yet construction spending there is down nearly 24% compared with 2025.

Source: futurism.com

The source text links the chart to Thompson, best known for his “abundance” theory: a political program built around deregulation and associated with neoliberal economic thought.

The durability question

For now, if the US economy has an abundance, AI appears to be supplying it. But that leaves a narrow foundation for a construction boom: data centers are soaring while most other categories contract, and investors are increasingly questioning whether AI spending can stay at its current level.

I think the chart’s most important signal is not simply that data centers are growing fast. It is how much the broader construction picture depends on that growth. If AI investment cools, the sector currently carrying the boom may stop doing so just as the rest of construction remains weak.

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X