- Oracle, Microsoft, Amazon, Meta and Alphabet plan about $800 billion in capital spending in 2026 as artificial intelligence infrastructure supports U.S. economic growth.
- Economists warn that the buildout could constrain investment in housing, industry and small businesses by raising borrowing, land, material and labor costs.
- The United States has 543 data centers under construction and another 2,513 facilities in the planning stage.
Rapid growth in artificial intelligence infrastructure is supporting U.S. economic expansion but could limit investment in other sectors, The Atlantic reported, citing economists and industry data. The planned 2026 capital expenditures by Oracle, Microsoft, Amazon, Meta and Alphabet are eight times their level a decade ago.
The spending is already visible in economic data. In the second quarter, private-sector investment excluding real estate increased at a 12.7% annual rate, while equipment purchases rose 21%. Spending on computers contributed 0.7 percentage points to U.S. gross domestic product growth.
Data centers compete for capital and resources
Economists point to a crowding-out effect as technology companies’ demand raises the cost of borrowed capital, land, construction materials and labor. Projects in other sectors may consequently become less attractive to investors.
The shift is particularly visible in construction. Spending on data center construction rose 65% from a year earlier, while investment in residential construction and retail properties fell 10%. Spending on educational facilities declined 13%, office buildings 14%, hotels and resorts 15%, and industrial facilities 24%.
Dallas illustrates the trend. About 125 data centers operate in the area, with another 19 under construction and 53 planned. Despite the investment inflow and persistently low unemployment, housing starts in the region fell 12% from a year earlier.
Economists also note that data centers provide limited long-term benefits to local communities. After construction ends, the facilities consume large amounts of electricity and water while creating relatively few permanent jobs.
Economic impact depends on AI productivity
Assessments of the AI boom’s effects remain mixed. Stanford University economist Ryan Cummings said rising costs are already pushing mining businesses out and described that outcome as positive.
Companies in other industries, however, may face more expensive loans and restricted access to resources. According to the experts surveyed, the overall economic effect will depend on how much artificial intelligence increases productivity.
Previous reporting said SWI Group raised about $900 million to upgrade and expand data centers in the United States and Sweden by repurposing former FTX mining assets for AI infrastructure.
Source: Incrypted
