June 2026 Completed New Home Inventory: 118,000

June 2026 completed new home inventory was an estimated 118,000, or 24.3% of new single-family homes for sale, seasonally adjusted. The rest were under construction or not started. The preliminary figures in the Census Bureau and HUD’s July 24 release put total supply at 9.3 months, a measure that includes all three stages.

For builders and agents assessing competition for buyers with near-term move dates, completed homes are the starting point. Even that category, however, does not guarantee immediate occupancy.

June inventory by construction stage

Seasonally adjusted homes for sale at month-end; Table 3a.

Completed inventory matched its year-earlier estimate even though the estimated total was lower. That distinction matters when comparing housing inventory reports: a lower total does not necessarily mean fewer finished homes competing for buyers.

These are month-end inventory counts, not annualized figures. Seasonal adjustment removes recurring seasonal patterns to make monthly comparisons more useful. Comparisons between construction stages should use the same adjustment basis.

Which homes can meet a near-term move date?

Under Census definitions, a house is generally completed when its finished flooring is installed; earlier occupancy also qualifies. Under construction means work has started but the house is unfinished. Not-started inventory consists of permitted homes awaiting construction.

For a buyer with a firm move date, check each completed home’s occupancy approvals, remaining work and builder closing schedule. The statistical classification alone cannot establish whether a property will meet that deadline.

Under-construction homes require a closer look at expected delivery dates. A nearly finished house could compete for the same buyer as a completed home, while one whose foundation work has just begun may not. Census’s combined category does not reveal that distinction. Not-started homes belong in the longer delivery pipeline unless the builder can substantiate a schedule that fits the buyer.

What Census counts as a new-home sale

The sales series covers only part of new construction activity: homes built for sale. Owner-built homes, contractor-built homes on an owner’s land and homes built for rent fall outside its scope, as Census’s construction-versus-sales comparison explains.

Census records a sale when a contract is signed or a deposit accepted. A house can be sold before construction starts, during construction or after completion. Sales therefore measure purchase commitments rather than completed closings.

A finished house already under contract is consequently outside the reported unsold inventory. Counting every house awaiting closing as an available competitor would overstate buyers’ choices.

Cancellations complicate that picture. After recording the initial sale, Census stops asking about sale status. It does not adjust sales for subsequent cancellations or count a later resale. The survey can therefore differ from a builder’s current availability list. Check cancellations and homes returned to market when assembling a community-level comparison.

How the sales pace produces 9.3 months of supply

Months of supply divides inventory by monthly sales. When sales are expressed at a seasonally adjusted annual rate, divide that rate by 12 first. As Census’s definitions explain, annualization describes the month’s pace; it is not a forecast of the year’s sales.

The June release reported an annualized sales pace of 628,000 and inventory of 485,000:

485,000 ÷ (628,000 ÷ 12) ≈ 9.3 months.

Holding inventory constant, a slower sales pace raises the ratio. Months of supply can therefore rise without any increase in the number of homes buyers can occupy soon.

Dividing completed inventory by that same all-stage sales pace gives approximately 2.3 months, our calculation using the rounded published figures. This measures completed stock’s contribution to the headline ratio. It is not a separate Census estimate of how quickly completed homes will sell: the denominator includes sales at multiple stages, and construction can replenish the completed category.

Preliminary estimates and statistical uncertainty

June’s estimates are preliminary. Census revises the preceding three months with each new sales release, which is why the table uses May as revised in July. Sampling and nonsampling errors, including incomplete or inaccurate responses, also affect the estimates.

June’s estimated monthly sales change was +1.6%, with a 90% confidence interval of ±14.8 percentage points, according to the release’s explanatory notes. That is insufficient evidence to establish an increase.

The construction-stage counts are estimates too. A small difference between rounded monthly figures does not, by itself, establish a trend in finished-home availability.

Regional totals cannot establish a community’s excess supply

Census’s sample supports national and four-region estimates, but is too small to produce state or local estimates. A regional total therefore cannot establish whether a particular subdivision has excess inventory.

Use broader homebuilding analysis as context for comparing available homes in the same price range and buyer market. Separate completed homes from those deliverable within the relevant move window, then examine recent contracts, cancellations and incentives. A concentration of unsold finished homes competing for the same buyers would warrant investigating pricing pressure.