United States · Land
The lot shortage ended. Most people are still pricing like it didn’t.
Finished-lot supply in the US has loosened for seven straight quarters. What that does to a land deal you underwrote in 2022.
For four years the answer to almost any land question in the United States was the same: there are not enough finished lots. Builders would take anything with a recorded plat and utilities to the stick. You could underwrite sloppily and be rescued by scarcity.
That is over, and the change is not subtle.
Zonda’s index rose for a seventh consecutive quarter in Q2 2026, and lot supply loosened in 27 of the 30 major metros it tracks over the preceding twelve months. Austin, Atlanta and Denver now read as significantly oversupplied. Los Angeles and Philadelphia still do not — which is the whole point.
The national number is the least useful number
An index that aggregates thirty metros tells you the weather, not whether it is raining on your parcel. A deal in an oversupplied metro and a deal in a constrained one are no longer the same trade, and they have not been for about eighteen months.
What matters on a specific site is narrower: how many competing finished lots sit inside the same school attendance zone and the same price band, who controls them, and how long the builders buying there have been extending their takedown schedules.
Scarcity was doing work in your model that you may have credited to your own judgement.
What this does to the numbers
Three things move at once, and they compound:
- Absorption slows. A sell-out you modelled at 24 months is the assumption most likely to break, and it is the one with the largest effect on a levered return.
- Takedowns stretch. Builders who competed for contracts now negotiate them, and the schedule is the first term they reopen.
- Your carry runs anyway. Interest, taxes and management do not care about absorption.
Meanwhile the cost side has not loosened with the supply. Impact fees, permitting and infrastructure requirements are embedded in the basis, and they do not come back out when the market softens.
What we would actually check
Before anything else, the downside case at 36 months rather than 24. If the deal only works at the base case, it is not a deal, it is a bet on a market that has already turned once.
Then the competing pipeline — not listings, but entitled-and-unbuilt lots, which is where the real overhang sits. Then whether the builders in that submarket are still signing contracts or quietly running out standing inventory at nine-plus months of supply, which is roughly where new single-family sat nationally in mid-2026.
None of this says do not buy land. It says the thing that used to cover a thin underwrite is gone, so the underwrite has to carry the deal on its own.
Sources