Spending on U.S. data center construction rose to a seasonally adjusted annual rate above $75 billion in July, nearly 60% higher than a year earlier, as artificial intelligence companies and cloud providers continued expanding computing infrastructure.
By yourNEWS Media Newsroom
The artificial intelligence infrastructure boom pushed U.S. data center construction spending to another major increase in July, with expenditures climbing nearly 60% from the same month a year earlier.
Spending reached a seasonally adjusted annual rate of more than $75 billion, according to Census Bureau data released Tuesday and reported by Axios.
The increase from July 2025 reflects the continuing race to construct the enormous facilities needed to house the computing capacity behind artificial intelligence systems and other data-intensive technologies.
Data centers typically resemble large warehouses and require substantial investment in land development, electrical systems, cooling infrastructure and specialized construction.
The Census Bureau’s construction-spending measure includes expenses associated with erecting those buildings, including labor, construction materials, engineering work, contractor profits, interest and taxes.
It does not include much of the expensive technology installed after construction is completed.
Servers, processors, memory chips, equipment racks and fiber-optic connections are excluded from the construction figure even though those components represent the majority of the total investment required to bring a modern data center online.
Commercial real estate firm Cushman & Wakefield estimates that construction itself accounts for only about 20% of a data center’s overall cost.
That distinction means the more than $75 billion annualized construction rate captures only one portion of the broader capital investment flowing into the AI and cloud-computing infrastructure buildout.
The July increase also suggests developers continued accelerating new projects despite growing political scrutiny over the resources required to operate the facilities.
Large-scale data centers can consume substantial amounts of electricity, particularly when packed with the high-performance processors used to train and operate advanced AI models.
Cooling those systems can also require significant amounts of water, depending on the technology and design of the facility.
Those demands have generated bipartisan concerns in communities and state governments confronting rising electricity needs, pressure on power grids and questions about whether local water supplies can accommodate rapid data center expansion.
Despite those concerns, construction spending continued climbing sharply through the summer.
The nearly 60% year-over-year increase indicates that companies remain committed to expanding physical computing capacity as competition in artificial intelligence drives demand for additional processors, storage, networking equipment and the buildings needed to support them.
With construction representing only a fraction of total data center investment, the Census Bureau figures provide another indication of the scale of capital now being committed to the infrastructure underlying the AI industry.
Original article
Comments