September’s nonresidential construction gains depended heavily on two connected markets: data centers and power. Without them, spending would have fallen, leaving the headline growth far less reassuring for contractors and property owners tracking demand across the wider sector. The distinction matters because a sector wide total can rise even when many of the businesses and project types within it are seeing weaker conditions. A positive monthly figure, on its own, doesn’t show how broadly construction demand is improving.
Data centers carry the increase
Private nonresidential construction spending rose 1.1% from August, the U.S. Census Bureau reported Monday. Data center spending climbed 8.6% over the same period, according to Associated Builders and Contractors’ analysis. Since May, spending across the category has risen at a 217% annualized rate. That last figure describes the pace implied by a short run of spending changes; it shouldn’t be read as a forecast that the same speed will continue.
The sector’s strength helped produce a sixth consecutive monthly increase in nonresidential spending. ABC chief economist Anirban Basu said data centers accounted for virtually all of September’s increase. “Excluding data centers and the closely related power category, nonresidential spending actually fell in September.” The qualification changes the meaning of the headline: the overall increase reflects a powerful pocket of investment, while other parts of the market collectively moved in the opposite direction.
Data centers also have a close relationship with power construction. Facilities built to store and process digital information need substantial electrical capacity, so demand in one category can support work in the other. That connection helps explain why the two are discussed together, though it doesn’t mean every power project serves a data center or that their spending will rise in lockstep.
A narrow base for growth
The instinct is to read another monthly increase as broad construction momentum. The breakdown argues for caution: much of the growth is concentrated in infrastructure serving the AI buildout, while spending elsewhere weakened. Basu expects that buildout to remain a major source of growth, but September’s figures show how dependent the overall total has become on a small group of projects.
This is more significant than it looks. When a few project categories account for most of an increase, the headline can give contractors and property owners a rosier impression of demand than the underlying mix supports. The reported gain is real, but it doesn’t establish that activity is strengthening across offices, retail, manufacturing, or other forms of nonresidential construction. The figures provided here don’t detail those categories’ individual performance, so the broad weakness shouldn’t be mistaken for a uniform decline in every market.
There’s another distinction between a surge in spending and a durable pipeline of work. Large facilities can involve substantial construction activity, and their scale can move an aggregate measure noticeably. But a single month’s data can’t tell readers how long current projects will continue, how many future projects will proceed, or whether other categories will recover. A sharp recent pace is evidence of strength in the measured period. Not proof of a lasting trend.
What the figures mean for the wider market
For industry professionals, the signal to watch is whether data center and power investment continues to offset declines across other nonresidential work. Until the gains spread, headline growth will say more about those two sectors than about the health of the broader market. If you’re working in this space, the useful question isn’t simply whether spending rose; it’s which kinds of projects are producing the increase and whether that concentration is changing.
Here’s the thing: narrow growth can still support meaningful business activity for firms positioned to serve these projects. But it offers less reassurance to companies whose work depends on other types of construction. The distinction matters in planning workloads, reading market reports, and judging whether a strong aggregate number reflects a broad upswing or a concentrated investment cycle (and this is the part most people overlook).
That said, September’s figures don’t settle what comes next. Basu’s expectation that the data center buildout will remain a major growth source points to continued support from that segment, but the available figures don’t establish whether it can keep offsetting weakness elsewhere. For now, the increase is best read as a sector specific gain with wider implications that remain uncertain. The headline is positive; the breadth is not.
U.S. Census Bureau construction spending report · ABC analysis
