The AI infrastructure buildout is no longer a tech story. It is the biggest economic bet in US history, and it is already competing with every other commercial real estate project for the same four things: land, power, labor, and capital.
Here is what the numbers actually say, and what Midwest CRE operators need to understand right now.
Projected US investment in AI infrastructure from 2025 through 2032: $10.3 trillion. That averages 3.6% of GDP annually. To put that in context, the railroad buildout peaked at 2.24% of GDP. The interstate highway system hit 1.13%. This is not a sector. This is a structural reorganization of where American capital goes for the next decade.
Alphabet, Amazon, Meta, Microsoft, and Oracle alone are projected to spend $4.2 trillion in the four years ending 2029. A growing share of that is financed off-balance-sheet through private credit and bank debt, which means the true leverage in this trade is not fully visible in any public filing.
Through July of this year, private data center construction spending hit a seasonally adjusted $37 billion, about $9 billion above the same period in 2025. Over that same window, all other private construction spending fell $46 billion below year-earlier levels. Apartments. Retail. Office. Industrial. All pulling back. Data centers: accelerating.
That divergence is not temporary. It reflects where capital is going and where lenders are comfortable underwriting.
This is the part most brokers are not paying attention to.
In Mississippi, a proposed aluminum smelter with 1,000 permanent jobs selected Oklahoma over Vicksburg because a nearby data center project consumed the available electricity. A thousand jobs went somewhere else because a data center needed the power first. That is not a hypothetical. That is a site-selection outcome happening right now.
The Federal Reserve Bank of Richmond reported that data center construction is straining labor availability across its region. Unionized electricians in the DC area went from 9,000 to 17,500 in recent years, driven almost entirely by data center demand. Equipment prices for computers, peripherals, and semiconductors are up 20% year over year, partly from data center chip demand bleeding into the broader supply chain.
Land costs near data center clusters are rising. Site-selection consultant Didi Caldwell confirmed data centers are pushing up land costs in targeted markets. If you own land near power infrastructure in secondary markets, that story is not lost on the institutional capital looking to deploy.
LinkedIn estimates AI created more than 750,000 new US jobs from 2023 through mid-2026. Data centers alone added 117,000 jobs since the start of 2024, not counting construction. The median salary for AI-related job listings is $180,000 versus $80,000 across all listings. That income concentration is already showing up in luxury real estate demand in Silicon Valley, and it will show up in Midwest gateway markets as remote and distributed work patterns follow the talent.
Kansas City, Tulsa, Omaha, and Des Moines are actively in the data center conversation. Available land, power infrastructure, lower land costs, and central fiber routes make secondary Midwest markets a genuine alternative to Virginia, Texas, and Arizona, which are hitting power and land constraints.
The opportunity is not just in owning or brokering data center sites. It is in understanding that the secondary effects of this buildout, tighter labor, higher land costs near substations, industrial site constraints, capital flowing away from traditional construction, are already reshaping how every other commercial asset gets valued and financed.
If AI revenue does not materialize fast enough to support the debt load being raised, losses will transmit through banks and private credit firms that funded this buildout. Off-balance-sheet structures make that exposure hard to quantify. The Fed and Goldman Sachs are already flagging it. This does not mean the buildout stops. It means the financing side deserves the same scrutiny as the demand side.
First, if you have industrial land within proximity to power infrastructure in the Midwest, get a data center site evaluation done. You may not know what you have.
Second, if you are underwriting industrial or flex deals near data center clusters, your labor and construction cost assumptions need to reflect the new competition for workers and materials. Your proforma from 18 months ago is stale.
Third, if you are advising clients on 1031 exchanges or portfolio repositioning, data center adjacency and power infrastructure access are now legitimate value drivers worth underwriting, not just talking points.
The capital is moving. The question is whether you are positioned where it is going.
Logan Freeman Managing Broker, Midwest CRE Advisors $500M+ closed across the Midwest 573-694-9669 | mwcreadvisors.com
Copyright 2026, Midwest CRE Advisors
