What It Costs to Build Self-Storage in 2026 (Per Square Foot)

Drive-up vs. climate, shell vs. all-in, and the two 2025 cost drivers most guides still leave out. Then the part that actually matters: turning $/sf into a go-or-no-go.

The most-searched number in this business is a simple one: what does it cost to build self-storage, per square foot? It is also the most misused — not because the figures are hard to find, but because almost everyone quotes the wrong kind of number, then drops it into a pro forma that's missing half the cost.

Here are the real 2026 ranges, what they include, the cost drivers that moved this year, and — the part contractor pages skip — how to take a cost per foot and turn it into the only answer that matters: does the deal pencil.

Shell, vertical, all-in: know which number you're holding

The single biggest mistake is treating a quoted $/sf as your total cost. It almost never is. There are three very different numbers:

• Building shell — the steel package and erection. The cheapest, most-quoted figure.
• Vertical construction — a fuller build: shell plus interior partitions, doors, basic systems.
• All-in — vertical plus land, site work, soft costs (architecture, engineering, permits, impact fees), contingency, and financing. This is your real cost.

Quote a shell number as if it were all-in and you'll underwrite a project that's 40%+ light. Here's the 2026 landscape, drawing on Storable's late-2025 cost guide and contractor data:

*All-in figures are syntheses that swing hard on land price — treat them as ranges, not a quote.

Climate-controlled costs roughly double the drive-up shell because you're adding insulation, HVAC, vapor barriers, and interior corridors. You build climate when land is expensive enough that going vertical and renting at a premium beats spreading out — not because it's "better." It's a different cost-and-revenue machine.

The two 2025–26 cost drivers most guides miss

If your cost basis came from a 2023 spreadsheet, it's already wrong in two specific ways:

1. Steel is up 9–14% year over year. Light-gauge framing — the bones of every storage building — climbed on Section 232 tariffs and tight domestic mill schedules. On a drive-up building that's almost entirely steel, that flows nearly straight to your hard-cost line.

2. The A2L refrigerant transition adds 6–10% to mechanical cost on climate-controlled product. The EPA's AIM Act phase-down changed the refrigerants new HVAC systems use, and the new equipment costs more. Drive-up dodges this; multi-story climate eats it.

What else moves the number

• Scale. Under ~40,000 sf, you're inefficient — fixed costs (office, sprinkler riser, site entry) spread over too little rentable area. The efficient zone starts around 50,000+ sf.
• Efficiency ratio. Cost is per gross foot, but you only collect rent on net rentable feet. Single-story runs 80–88% efficient; multi-story with elevators 65–75%. A ramp/over-under multi-story design claws that back to ~88–90%. The lower your efficiency, the more your real cost-per-rentable-foot balloons above the headline.
• Site & jurisdiction. Soil conditions, stormwater retention, mandated façades in urban infill, sprinkler requirements, and elevators all swing the number.
• Land. The wild card. The old "land is 25–30% of cost" rule is no longer reliable — land has become the variable that makes or breaks deals (more on that below).

From $/sf to a real answer: yield on cost

Here's what no contractor cost page will tell you: a cost per square foot, by itself, tells you nothing about whether to build. A $90/sf deal can be a home run and a $70/sf deal can be a disaster. What matters is cost relative to the income the building will throw off.

The bridge is yield on cost — stabilized net operating income divided by total project cost — measured against the cap rate you'll exit at. You want yield on cost to clear the exit cap by a healthy margin (a 200-basis-point development spread is the common bar). That gap is your profit and your cushion.

So the cost number only becomes useful once you run it through the chain: gross area → efficiency → rentable feet → rent → NOI → yield on cost → spread. A higher cost is fine if rents justify it; a low cost is worthless if the land was priced for apartments. Land priced for a higher-and-better use is the most common way a storage deal with a "reasonable" $/sf still fails — because storage is a low-rent-density use and can't carry an expensive basis.