Data center construction more than doubled in two years to an $85 billion pace. New home building fell.

Rows of server racks lining a data center hall

Key points

  • Data center construction spending reaches a record pace
  • New home construction spending falls over the same two years
  • Data centers account for most private office construction

Private spending on U.S. data center construction more than doubled in two years, reaching a record annualized pace of $85 billion in August. Spending on new houses and apartments moved in the opposite direction, with its annualized rate falling by $29.2 billion over the same period.

The data center figure rose from $36.7 billion in August 2024 and is the highest in the Census Bureau's data center series, which begins in January 2014, according to data released Oct. 1. The figures are seasonally adjusted spending rates, not amounts spent in a single month.

Data centers grow as housing construction falls

The table compares private construction spending in August of each year, expressed as annualized rates in billions of dollars, from the Census Bureau's monthly series.

CategoryAug. 2024Aug. 2025Aug. 2026
Data centers$36.7$49.1$85.0
New single-family homes$421.9$418.1$403.3
New multifamily$126.4$116.5$115.8
Total private construction excluding data centers$1,693.3$1,658.3$1,570.3

Data center spending rose 73% from a year earlier. Over two years, private construction excluding data centers fell 7.3%. The annualized spending rate for new single-family homes dropped by $18.5 billion, while new multifamily construction fell by $10.7 billion. Changes are calculated from unrounded figures.

Data centers now drive office construction

The Census Bureau counts data centers as part of its office category. Private office construction spending rose 29.8% from a year earlier, to $134.1 billion in August. Data centers made up $85.0 billion of that, while spending on general office buildings slipped to $45.8 billion from $50.7 billion a year earlier.

Data centers accounted for 11% of private nonresidential construction spending in August. Two years earlier, the share was 4.6%.

The bigger estimate behind the headlines

Stateline described the data center buildout as the largest financial bet in U.S. history, citing a Brookings paper by Columbia Business School professor Stijn Van Nieuwerburgh. The paper projects $10.3 trillion of investment in data centers and related AI infrastructure from 2025 to 2032. We covered that estimate in September, when a Fed president asked whether AI is becoming too big to fail.

The Census figures are narrower. They measure construction put in place and, by the bureau's definitions, exclude racks and servers in data centers. They also aren't adjusted for inflation. The August numbers are preliminary, and the Census Bureau will revise them with its next release on Nov. 2.

Frequently asked questions

How much is the U.S. spending on data center construction?

Private data center construction spending ran at a seasonally adjusted annual rate of $85.0 billion in August 2026, according to Census Bureau data released October 1, 2026. That was up from $49.1 billion a year earlier and $36.7 billion in August 2024, and it is the highest in a series that begins in 2014.

Is data center construction crowding out housing?

Over the two years to August 2026, data center construction spending rose by $48.3 billion while spending on new single-family homes and apartments fell by $29.2 billion, Census data show. Private construction outside data centers fell 7.3% over the same period. The data show the two moved in opposite directions but do not prove one caused the other.

Does Census construction spending include servers and chips?

No. The Census Bureau measures the value of construction put in place, and its definitions exclude racks and servers in data centers. The figures are also not adjusted for inflation, and the August 2026 numbers are preliminary until the November 2 release.

More coverage

Latest stories

Dennis Singleton
Dennis Singleton

Dennis Singleton was born in Australia and later moved to the United States. He has spent years following the markets, but what keeps his attention is how AI is built. He writes about the companies behind the technology, from semiconductor designers and advanced packaging to photonics, memory, networking, and the hardware powering modern AI. His approach starts with filings, earnings, and industry research, then translates the important details into clear, straightforward analysis without unnecessary hype.