May multifamily starts (buildings of 5+ units) plunged 41.6% to a 284,000 seasonally adjusted annual rate — the lowest since November 2024 and, measured month-over-month, the steepest single-month drop since April 2009. Total housing starts fell 15.4% to 1.177 million, the weakest reading since May 2020.

The pullback was regional and broad. Starts fell 17.0% in the South and 17.2% in the West — the two regions that absorbed the 2022–2024 Sun Belt construction wave — alongside a 26.8% drop in the Northeast. Only the Midwest rose, up 3.7%.

The cause is unglamorous: high financing costs, construction inputs up roughly 9.6% year-over-year, and developers refusing to break ground into markets still digesting concessions. The result is a pipeline switching off in exactly the metros that have been most oversupplied.

Here is the part the headline misses. A starts collapse is bad news today and a setup tomorrow. Apartments take 18–30 months to deliver. Units not started in mid-2026 are units not delivered in 2027–28 — precisely when today’s oversupply will have leased up.

That creates a supply air-pocket. The Sun Belt markets posting the deepest rent resets and the steepest value declines right now — workforce and Class C assets down 20–30% in the worst cases — are the same markets whose future competition just evaporated. The pain is real; the forward setup is the opposite of the pain.

This is the divergence Altus is measuring from the pricing side: a record median price of $129/sf set by smaller, well-located assets that can actually trade, while the broad market stays frozen. Record prices on shrinking buildings and a collapsing pipeline are two readings of the same instrument — capital concentrating into what survives the reset and positioning for what comes after it.

The discipline is to separate the trough from the trend. A deal underwritten to mid-2026’s soft rents, with no credit for the 2027–28 supply vacuum, is buying the bottom and pricing the bottom as permanent. That is where capital gets protected — or destroyed.

Developers: the window to deliver into the air-pocket is now closing — anything not breaking ground this year misses the 2027–28 demand-supply gap. Owners/operators: the concession war has a visible end date; hold-and-renew logic strengthens. Lenders: forward supply risk in the Sun Belt is falling even as today’s distress prints — a reason to underwrite stabilized exits more confidently than spot rents suggest. Buyers: the cleanest entries are in markets where current softness and future scarcity overlap.