The second-quarter apartment data is the clearest sign yet that the multifamily cycle has inflected. The U.S. absorbed more than 187,000 units from April through June — a pace well above the seasonal norm — and national occupancy climbed back to 95.5%. It was the sixth consecutive quarter of falling annual supply.

The engine of that turn sits on the supply side. Annual deliveries have dropped to roughly 340,200 units in the year ending Q2, down from a peak near 588,000 in late 2024, and now below the decade average for the first time in three years. Only about 77,700 units delivered in the quarter itself. The wave that flooded lease-ups for two years is visibly receding.

Rents are following, but slowly and unevenly. Effective asking rents rose 1.4% in the quarter yet remain 0.2% below year-ago levels, and concessions were still offered at 24.6% of apartments. Regionally, the South is the only part of the country with annual rent declines and occupancy under 95% — it is still digesting its own construction boom.

The temptation is to call this a demand recovery. It isn't, quite. Trailing-year absorption of about 271,300 units is still below the decade average near 340,000. Demand is healthy, not booming. What actually turned the market is that the industry stopped delivering — the taps closed faster than the renters arrived.

That distinction is the whole underwriting story. A supply-led turn firms the numbers in a specific order: occupancy first, then effective rents, then asking rents and the burning-off of concessions. It also firms them unevenly — only in the submarkets where the pipeline genuinely emptied. This is not a rising tide. It's a low tide going out at different speeds in different harbors.

So the market splits by delivery schedule. Much of the Midwest and Northeast — where supply was never as heavy — is already tightening, and owners there can underwrite firming occupancy today and rent power into 2027. The Sun Belt is a different market: still absorbing elevated deliveries, still paying concessions, still posting sub-95% occupancy. Underwriting the "national recovery" as one number blends two markets moving on different clocks.

There's a forward tension worth holding, too. Deliveries are collapsing because starts collapsed a year ago — and starts collapsed because construction financing dried up. Yet capital is selectively funding the next wave: a lender just wrote its largest-ever construction loan for 530 new luxury units in Northern California. The developers who can get financed now are aiming at the 2027–28 supply trough, which is exactly when this turn should be tightest. Scarcity of financing today is scarcity of competition tomorrow.

Whether rent growth broadens from occupancy is the open question. Occupancy at 95.5% is a firming signal, but effective rents are still negative year-over-year and concessions remain elevated. If the supply drop keeps compounding into 2027, pricing power returns to landlords broadly; if demand softens with the labor market, the turn stalls at occupancy and never reaches rents. Q2 REIT prints landing in late July will be the first read on which way it breaks.

The apartment market turned — but it was the construction slowdown, not a demand boom, that did it. Underwrite the delivery schedule, submarket by submarket.