The Constraint Has Moved

For most (really all) of my professional career, the question that decided whether a deal happened was whether capital would show up and at what price. However, that is no longer the primary binding constraint. Capital may not be as attainable with higher rates, but it is generally available. What is scarce now is physical: dirt that is zoned for the use you need, electricity that can actually be delivered to the meter, and in the markets that overbuilt, the simple passage of time required to absorb what already got built.

Three major insights from the last few months make the same point from different lenses of perspective. Industrial outdoor storage, a property type that barely had a name five years ago, is repricing because the AI buildout needs somewhere to stage steel and transformers and there is nowhere left to put them. Site selectors have reordered their entire decision framework around grid capacity, displacing labor and tax incentives at the top of the list. And the Sun Belt, which spent four years absorbing a historic wave of apartment supply, is finally working through it at three distinctly different speeds depending on local absorption.

The Dirt Behind the Data

The most interesting trade in the AI buildout may not be the data centers themselves but rather the gravel lots next to them. Industrial outdoor storage has gone from an overlooked, largely unloved property type to one of the better performing corners of commercial real estate, and the arithmetic behind it is straightforward. DAT Freight and Analytics estimates that every gigawatt of new data center capacity requires roughly 100,000 truckloads of materials. All of that steel, pipe, cable, and equipment has to be staged somewhere within reasonable distance of the site, and staging requires land that is zoned to allow it.

That land is getting harder to find, not easier. New IOS sites run into restrictive zoning and consistent local opposition, which means supply barely grows. At the same time, data centers are consuming the existing inventory. In Manassas, Virginia, more than 200 acres currently used for outdoor storage are slated to be converted into data centers, which would eliminate more than twenty percent of local supply in a single stroke. The use that is driving demand is also destroying the supply that serves it.

The rent numbers reflect that squeeze. CBRE puts nationwide IOS rent growth at thirty-two percent since 2021, and across eight major markets, including Houston, Savannah, and Norfolk, rents have risen more than 110 percent since 2020. The total U.S. IOS market is now estimated at roughly $300 billion. Pricing power sits with the landlord in a way it rarely does in industrial, partly because a data center contractor spending hundreds of millions on a campus can absorb a rent increase on a laydown yard without blinking. Storage is a rounding error in their budget and a critical path item in their schedule, which is close to an ideal position for an owner.

Institutional capital has noticed. Brookfield announced a $1.2 billion acquisition of Peakstone Realty Trust. J.P. Morgan partnered with Zenith IOS to assemble roughly 100 properties across 33 markets. Catalyst Investment Partners has built a $1.4 billion portfolio, and Blackstone has financed more than $1 billion in IOS transactions. The sector is still overwhelmingly owned by small, local operators, which is exactly the fragmentation profile that invites aggregation. Infrastructure construction, supported by a healthy municipal bond market, is adding a second tenant base beyond the AI story.

For owners of marginal industrial land, functionally obsolete flex buildings, or sites near a transmission corridor, this is worth a hard look. The asset that does not pencil as a building may pencil quite well as a yard.

Power Is the New Location

The oldest cliche in this business is about to get an amendment. Electricity has overtaken land, labor, and incentives as the single most important factor determining where capital-intensive projects get built in the United States, and the shift happened faster than almost anyone expected.

Among the 72 members of the 2026 Site Selectors Guild, 79 percent ranked power access as a top influence this year, with labor and geopolitics tied well behind at 50 percent each. More than half said inadequate power supply alone is now sufficient grounds to eliminate a location from consideration before any other analysis begins. Compare that to the prior two surveys. Tariff policy topped the 2025 list at 87 percent, with utility capacity drawing 66 percent, and as recently as 2024 workforce availability ranked first at 76 percent. The grid jumped from a secondary consideration to the primary screen in roughly twenty-four months.

The supply and demand picture explains why. Goldman Sachs Research projects U.S. electricity demand from capital-intensive projects rising from 31 gigawatts in 2025 to 66 gigawatts by 2027, a more than doubling led by data centers. Utilities cannot build that fast, and equipment lead times, particularly for transformers and switchgear, have become the gating factor rather than capital or permitting.

Community opposition has become a screening factor in its own right. Twenty-two percent of consultants cited local resistance as grounds to drop a site outright, and Data Center Watch counted at least 45 projects worth roughly $68 billion blocked or delayed between April and June, following 75 projects worth $130 billion in the first quarter; a meaningful share of the pipeline failing at the local level.

The practical implication for owners, developers, and brokers is that documented power capacity is now a marketing asset. Sites with secured interconnection win on both price and timeline, and the groups that gather and certify that data before going to market are capturing a real premium. Lining up power before entitlements has become the clearest path to winning the next capital-intensive deal. If you own land near substation capacity, the value of that proximity is worth quantifying now rather than discovering it at the closing table.

The Sun Belt Clears Its Glut at Three Speeds

The Sun Belt's apartment oversupply is finally working itself off, but the recovery is uneven enough that treating the region as one market will produce bad decisions. The pandemic boom added 673,000 apartments and pushed rents up thirty percent between 2021 and 2024. Yardi now expects the major oversupplied markets to return to positive rent growth by the end of 2027.

The supply side is correcting hard. Sun Belt construction starts have fallen seventy-five percent from the 2022 peak. National deliveries are projected to drop from 696,000 units in 2024 to 421,000 in 2026, with another twenty-four percent decline expected in 2027. That is the kind of pullback that sets up the next rent cycle, and it is happening while capital remains cautious on the asset class.

The spread between markets is where the opportunity sits. Austin and San Antonio are leading the recovery, with absorption outpacing completions. Phoenix absorbed 17,000 units in the first half of 2026 alone. Dallas-Fort Worth, by contrast, still carries a pipeline exceeding 40,000 apartments, which will take time to clear regardless of how strong demand proves. Looking out further, Phoenix, Charlotte, Fort Worth, and Austin could face their tightest apartment conditions in years by 2028 if the construction slowdown holds.