Millrose's earnings report is a household forecast.
A household forecast is a storage demand forecast.

A public company owns the land under 143,771 future American homes across 30 states, and discloses the entire book every quarter. Almost nobody in self storage reads it. Here is the chain from a land option to a rented unit — and what is already standing where those rooftops land.

01 / FIRST PRINCIPLESFollow the household

Every forecast in this industry starts with population, or with rate trends, or with what the REITs said last quarter. Those are all downstream. Work the problem from the bottom and there is only one input that matters, and it is countable years in advance.

Now put a storage clock next to it. Vertical construction runs roughly 12 to 18 months, inside a two-to-three-year cycle from entitlement to opening. Then lease-up: Extra Space defines a store as stabilized at three years, Public Storage says three to five, and Newmark reports brokers now seeing five to six in oversupplied markets. The households Millrose financed this quarter arrive in 2030. Work backwards and the decision is now.

02 / THE COMPANYWhy this book and not somebody else's

Millrose was spun out of Lennar on February 7, 2025 — a taxable spin-off of $5.5B of land and $1.0B of cash, one Millrose share for every two Lennar shares. The mechanics matter, so precisely: it takes title to residential land — sometimes bought from the builder, sometimes from a third-party seller the builder identified — and is “solely responsible for funding Land Development up to the predetermined budgets.” The builder performs the work; Millrose writes the cheque. Horizontal development is defined in the filing as “work performed by our counterparties relating to the installation of utilities and infrastructure.” The builder posts a non-refundable deposit and pays a monthly option fee for the right to buy those homesites later, on a pre-negotiated schedule. If the builder starts vertical construction before exercising, it carries that cost itself.

Three things make it the cleanest signal available. It is public, so the whole portfolio is disclosed quarterly, state by state. It is diversified — 19 counterparties now, with 32% of invested capital outside the founding Lennar agreement. And the schedule has held: zero option terminations since inception, across every parcel, every counterparty. Builders are not walking away from this dirt.

03 / WHERE THE HOUSEHOLDS GOThe map, in doors not dollars

Read this in homesites, not capital. A $264,000 lot in California and a $66,000 lot in Oklahoma each produce exactly one household that needs somewhere to put its things.

Two things jump out. Texas and Florida are 42% of the future rooftops. And Maryland runs 742 lots per community — 4,450 homesites in just six projects, by far the most concentrated in the book. Six locations in the Baltimore–Washington corridor are each going to drop a small town's worth of households into a single trade area.

04 / WHAT IS ALREADY THERERead this one chart and you have it

One bar per market. The line down the middle is exactly enough storage — the square footage those households need, from the SSA's own penetration rate. Solid is what is standing today. Faded is what is under construction. Anything past the line is a market that already has more storage than its households will absorb.

Why not just use square feet per capita?

Because it disagrees with itself. The same metric is published as 6.1, 7.4 and 7.8 depending on whose facility universe you count and whether the denominator is the whole country or just the metros a vendor tracks. Worse, it is blind to household size — and a storage unit is rented by a household, not a person.

Rank these twenty markets both ways and four of them move five places or more. The Inland Empire is the fifth-thinnest market on square feet per capita and the eleventh on a household basis, because its households average 3.40 people — the largest here. Phoenix moves the same six places. Tampa moves six the other way: it looks adequately supplied per capita and is the fifth most undersupplied per household.

Per-capita would have told you to build in the Inland Empire. The households say it is already 1.32× supplied, and 1.46× once the pipeline lands.

05 / PUT THEM SIDE BY SIDERooftops coming, storage standing

Two-thirds of the rooftops America's largest builders have committed to are landing in markets that will already have too much storage before those houses are framed. One-eighth are landing where they won't.

The three that clear are not the fast-growing markets. Baltimore is the slowest-growing market in the sample at +2.1% and Los Angeles is second at +2.6%; San Francisco sits mid-pack. What they share is not demand — it is that supply cannot respond. Under construction is 5.3% of standing inventory in Baltimore and 6.9% in San Francisco, against Jacksonville at 16.3% and Orlando at 15.7%. The same entitlement difficulty that makes a California lot cost $264,199 makes a California storage facility hard to build.

That is the whole read, and it points somewhere counterintuitive: the markets where builders are buying the most dirt are the markets a storage developer should be most careful in. Cheap land means easy approvals, and easy approvals mean somebody already built the storage. Draw your own conclusions about where that leaves capital.