The Glut Ends Early

For two years the apartment call was simple: wait for the supply to clear. It cleared. The problem is that the pipeline behind it is nearly empty.

The U.S. apartment market crossed the line the entire recovery thesis was built around. Over the trailing four quarters, demand exceeded new supply for the first time since early 2022 — and it happened while the construction pipeline fell to its thinnest level in more than a decade. Two days ago the honest read was that starts were collapsing on schedule but demand had wobbled. The Q2 data revises that read, and it is worth saying so plainly.

The Signal. Cushman & Wakefield's Q2 2026 U.S. Multifamily MarketBeat, released July 15, puts net absorption at 124,600 units for the quarter — the fifth-highest quarterly total in nearly 25 years and an 8% increase year over year. National vacancy declined 35 basis points to 8.9%, its first reading below 9% since 2024. On a trailing four-quarter basis, roughly 362,000 units were absorbed against roughly 358,000 delivered.

The supply side is the more durable half. Only 88,000 units delivered in Q2, down 27% year over year and more than 40% below the mid-2024 quarterly peak. Approximately 475,000 units remain under construction — 3.5% of existing inventory, the lowest share since 2013. National asking rents rose 1.5% year over year, accelerating from 1.1% in Q1, the first improvement in annual rent growth in about a year. San Francisco leads the rent recovery at 13.0%, followed by San Jose at 7.0%.

The geography is the surprise. Of the 20 markets that expanded inventory most since 2019, 18 recorded quarter-over-quarter vacancy declines, averaging nearly three times the national rate of improvement. Dallas/Fort Worth absorbed 18,600 units in the first half, Phoenix 17,000, Atlanta 13,300, Austin 13,200. New York led the country at roughly 19,500.

Our Read. The consensus underwriting for the last two years treated Sun Belt oversupply as a multi-year problem and gateway markets as the safe harbor. The Q2 data says the opposite is happening at the margin. The metros that took the most supply are clearing it fastest, precisely because supply was the constraint — not demand. Once deliveries slow in an oversupplied market with real population inflows, the absorption math turns quickly.

That has a direct consequence for anyone still underwriting a discount. The supply-driven entry point in high-delivery markets was always a window, not a condition. The crossover printing in Q2 means the window is measurably narrower than it was in January, and concession assumptions carried over from 2024 and 2025 are now stale inputs.

The more interesting risk is on the other side. A pipeline at 3.5% of inventory is not a normalization — it is an undershoot. Deliveries decline through 2027 while absorption runs near the strongest levels of the past 25 years. That sets up a genuine supply shortage in 2028 and beyond, and the development capital that would normally respond is still constrained by financing costs and construction inflation. The cure for the glut is becoming the cause of the next squeeze.

Worth naming the counterweight: absorption at this pace has to be supported by household formation, and job growth and population gains have both been moderating. Cushman & Wakefield flags that demand held up despite those headwinds — which is encouraging, but it is a resilience observation, not a guarantee. One strong quarter is a turn. It is not yet a trend.

Stakeholder lens. Owners in high-delivery Sun Belt markets should stop underwriting a long grind and start modeling the lease-up already showing in the vacancy data. Buyers hunting supply-driven discounts face a closing window and should price the crossover, not last year's conditions. Developers with sites and capital are looking at a 2028–29 delivery window into a structural shortage — the entitlement work done now is the option that matters. For lenders, vacancy and concession assumptions built on 2024–25 comps are stale; the deliveries schedule through 2027 is the more useful input. And for anyone reading this week's Chicago Loop conversion trade at roughly $65 per foot, note that adaptive reuse is quietly becoming one of the few residential supply channels that still pencils.