A Quarter of Operators Are Above 95%. A Quarter Are Below 80%. The Middle Is Gone.

The Middle Has Gone Missing

The most useful number published about senior living occupancy this month is not the average. It is the spread.

ASHA's 2025 State of Seniors Housing report found that many operators have settled into average occupancy between 89% and 91% — a healthy-sounding band that describes the sector's center of gravity. But roughly a quarter of surveyed operators reported average occupancy below 80%, and roughly the same share reported 95% or higher ([Senior Housing News, August 28, 2026](https://seniorhousingnews.com/2026/08/28/narrowing-capacity-little-growth-welcome-to-senior-livings-new-era/)).

That is not a normal distribution with a fat middle and thin tails. That is a barbell. Half the industry sits in the comfortable band; the other half is split almost evenly between buildings that are effectively full and buildings that are a quarter empty. Averaging them produces a number that describes almost nobody.

For anyone underwriting an acquisition, this matters more than the sector-level occupancy print. A market-average assumption applied to a specific building is now very likely to be wrong in one direction or the other, and the two errors are not symmetric. Underwriting census upside into a 95%-occupied asset is a modeling error with no recovery path. Underwriting it into a 78%-occupied asset may be the entire thesis.

The Ceiling Nobody Priced

The reason the top of that barbell is getting crowded is arithmetic.

NIC MAP reports that occupancy has crested 90% in primary markets and 93% across primary and secondary markets, and a NIC report released in late August found that more than three-quarters — 77% — of all senior living properties in primary markets are now operating above 85% occupancy ([Senior Housing News, August 28, 2026](https://seniorhousingnews.com/2026/08/28/narrowing-capacity-little-growth-welcome-to-senior-livings-new-era/)).

Against that, there were 16,000 units under construction in the second quarter of 2026 — against average absorption of roughly 32,000 filled units per year over the past four years (NIC MAP, via SHN, August 28, 2026). The industry is currently building half of what it absorbs annually, while three-quarters of its primary-market inventory is already north of 85% full.

The consequence is not a demand problem. It is a capacity ceiling. NIC MAP CEO Arick Morton put it directly: "If we can't keep up with demand growth, the natural consequence is a limit or cap on how far penetration can go … and it won't play out uniformly. Some markets are more undersaturated than others, so this is not a purely national story" (SHN, August 28, 2026).

That last clause is the operative one for a broker. "Not a purely national story" means the national occupancy figure has stopped being an underwriting input and become, at best, background. The question is whether the specific submarket has slack — and increasingly, most primary submarkets do not.

Full Is Not the Same as Finished

There is a temptation to read high occupancy as a solved problem. The operators living it do not.

NIC Senior Principal Omar Zahraoui was careful about this: strong occupancy "creates a favorable backdrop, but what it means for operating performance depends on what happens at the property level. That will vary considerably depending on their cost structure, labor efficiency, pricing, care delivery, other operating factors, and local market conditions" (SHN, August 28, 2026). He framed the sector's next challenge as "converting occupancy into sustainable operating margins while managing labor and other operating costs," and noted that communities approaching full occupancy "will need to manage demand and waitlists more strategically."

The conversion is achievable, and there is a published example of what it looks like. Beztak reported 94% occupancy across its All Seasons and Monark Grove brands with a 48.6% operating margin at the end of the first quarter of 2026 (reported April 2026, via SHN, August 28, 2026). Notably, Beztak budgets its communities to operate at 90% and manages costs against that assumption rather than against its actual census — the discipline is in refusing to spend the upside.

That is the behavioral difference between the top of the barbell and the middle. When census can no longer be the growth lever, the lever becomes expense structure, and the operators who were already running that way do not have to learn it under pressure.

The Asset That Still Has Room

The most interesting figure in the August 28 reporting is one that reads at first like a liability: more than a third — 40% — of existing U.S. senior living units sit in communities that are 25 years old (NIC MAP, via SHN, August 28, 2026). Some of that stock will cycle into functional obsolescence, which makes the supply gap worse than the construction numbers alone suggest.

But Morton's read on the same stock is the one worth underwriting against. Asked where capacity can actually come from, he pointed inward rather than to new development: "Everybody should be thinking about it: what can I add on land I already control, and what can I reposition inside the buildings I already run? That's the fastest capacity this industry can create. After that it's about getting to scale, figuring out where development is actually possible and who you can partner with to do it. And all of that work is local."

He named two specific sources. First, unit-level latent capacity: "double occupancy in AL and MC has been a feature of this industry for decades, so there's real latent capacity inside buildings that already exist." Second, dirt: "the blessing of a very old stock is that a lot of it sits on excess land — land was cheap and plentiful when most of it was built — so there's a good deal of room to expand on dirt operators already control" (SHN, August 28, 2026).

That reframes what an older community is. A 1998-vintage building on eight acres in a submarket running at 92% is not simply a dated asset with capex exposure. It is an entitled site with expansion optionality in a market where the replacement pipeline cannot compete inside a five-year hold.

What We Are Telling Clients

Three things follow for anyone buying or selling senior housing in this market.

Stop underwriting to the market average. Given a barbell distribution, the sector print tells you almost nothing about a specific asset. Get the actual trailing census by month, and get the waitlist. A building at 95% with a waitlist and a building at 95% without one are different assets.

Price the land, not just the building. If excess acreage on a controlled site is now the fastest route to new capacity, then parcel size, zoning and setback capacity on an existing campus belong in the valuation rather than in the appendix. This is the single most underweighted input we see in seller packages.

For sellers, the census story has an expiration date. An asset in the low 80s can still be sold on a credible lease-up thesis today, because a buyer can point to a submarket with room. As primary markets tighten past 85% and 90%, that argument gets harder to make and the discount widens. The window for selling occupancy upside is narrower than the window for selling in-place cash flow.

None of this depends on a demographic forecast. It depends on buildings that already exist, census that has already been achieved, and a construction pipeline that has already not been started.