Houston, Tampa, Phoenix and Cape Coral are still underwater while Minneapolis and Boston lead the recovery. History and the supply pipeline suggest timelines for the laggards, and they are not all the same.

Key takeaways

• Composite same-store NOI rose 5.2% in Minneapolis and 4.7% in Boston in the second quarter of 2026, while Houston fell 5.8%, Tampa fell 5.3% and Cape Coral-Fort Myers fell 14.3%.
• Recovery timing in an overbuilt storage market runs on three clocks: how much supply is still under construction, a roughly three-year lease-up for each new facility, and whether demand gets a shock.
• Rough sequencing: Dallas-Fort Worth in 2027, Tampa and Orlando behind it, Phoenix in 2028, Cape Coral at the end of the decade, and Houston cyclically sooner but structurally capped by inventory above 10 square feet per capita.
• A housing market thaw would compress every one of those timelines; renewed development in Sunbelt pipelines would reset them.

The self-storage recovery that operators described on second-quarter earnings calls is real, but it is not evenly distributed. Strip away the national averages and the sector splits into two countries. In one, markets that never overbuilt are already posting the kind of growth the industry has not seen since 2022: composite same-store NOI across the major REITs rose 5.2% in Minneapolis in the second quarter, 4.7% in Boston, 4.4% in Honolulu and 4.0% in New York. In the other, the development magnets of the 2021 to 2022 capital wave remain underwater, with Houston down 5.8%, Tampa down 5.3%, Dallas-Fort Worth down 4.9% and Orlando down 3.4%. Cape Coral-Fort Myers, the most extreme case in the country, posted NOI down 14.3%.

That divide has now held for four consecutive quarters, long enough that it can no longer be dismissed as noise. For the owner, lender or prospective buyer sitting in one of the lagging markets, the operative question is no longer whether the market is oversupplied. It is how long the hangover lasts, and the encouraging news is that this is a question with an approximate answer. Storage oversupply resolves on a knowable clock, driven by three variables: how much space is still coming, how long new space takes to fill, and whether demand does anything unusual in the meantime. Each of the lagging markets sits at a different point on that clock.

How the Sunbelt got here

The origin of the problem is familiar. Capital followed the pandemic migration maps, and between 2021 and 2022 developers concentrated an extraordinary share of the national pipeline in Sunbelt metros whose storage performance was then setting records. The resulting deliveries crested in 2023 and 2024, the two heaviest supply years in the industry’s history, and they landed just as the demand surge that justified them was fading.

The saturation shows up most clearly in per-capita inventory. StorageCafe’s 2026 supply analysis puts Houston, San Antonio and Jacksonville above 10 square feet of storage per resident, well above the 7 to 8 square feet long treated as rough national equilibrium, and finds rent declines “nearly universal” in markets past that threshold. The recovery’s leaders sit at the opposite pole: New York has about 4 square feet per capita, Boston 5, Los Angeles 5.1. Those constrained coastal markets are holding rents flat or pushing them up while Houston’s average street rate fell 3.3% over the past year to about $116 a month and Cape Coral-Fort Myers fell 5.8% to $129.

The divide is visible inside REIT disclosures too. When Extra Space Storage reported that new-customer rates had turned positive year over year in the second quarter, the markets it named were Austin, Dallas and Miami. The markets where they remained negative were Houston, Tampa and Phoenix. Same company, same platform, same quarter: the difference is supply.

Three clocks

Recovery timing in an overbuilt storage market is governed by arithmetic more than sentiment, and the arithmetic has three parts.

The first clock is the pipeline drain. Space under construction today is rate pressure tomorrow: Yardi Matrix data show average construction timelines have stretched to a series-high 431 days, which means a project breaking ground now delivers in late 2027, and everything already out of the ground delivers between now and then. Nationally that pipeline has shrunk to about 45 million square feet, or 2.2% of stock. The lagging markets are not all in the same position against that average. Phoenix still carries 6.9% of its existing stock under construction, roughly triple the national figure, and Sarasota-Cape Coral carries 5.4%. Houston and Dallas-Fort Worth, by contrast, have already worked their pipelines down toward the national norm; their problem is the supply that has landed, not the supply still coming.

The second clock is lease-up. A newly delivered facility typically takes about three years to reach stabilized occupancy, and for those three years it behaves as a discounter, buying customers with promotional rates that drag street pricing down for every competitor in its trade area. This is why rate pressure outlives the construction cranes. A market whose last big delivery cohort arrives in 2026 should expect that vintage to weigh on pricing into 2028 even if nothing else breaks ground.

The third clock is demand, and it is the wild card that history says matters most.

What the last cycle teaches

The industry has run this experiment before. Houston was the poster child of the previous supply cycle, absorbing a wave of deliveries that began in 2016 and did not meaningfully slow until 2019, and Texas broadly overbuilt into 2018. Street rates in those markets declined for roughly four years, a slog that felt permanent to the operators living through it. What ended it was not patience. It was the pandemic: a 12-month construction slowdown collided with the greatest demand surge in the sector’s history, and by mid-2021 Yardi Matrix analysis showed Texas street rates running 30% above 2018 levels with occupancy around 90% in every major metro.

The honest lesson from that episode cuts two ways. Absorption alone, at normal demand, healed the overbuilt markets slowly, on a timeline measured in three to five years from peak deliveries. A demand shock compressed the final stretch of that timeline into months. The post-financial-crisis period teaches the same lesson from the other direction: when demand recovered into a construction pipeline that credit markets had shut down, REIT occupancy climbed from the low 80s to the mid 90s between 2010 and 2015 and new space was absorbed in half the usual time.

Applied to today’s map, the framework says recovery dates depend on when each market’s deliveries peaked, what remains in its pipeline, and what demand does. On the first two variables the answers are known. On the third, the frozen housing market is the obvious candidate for a shock, and no forecaster currently has turnover normalizing before 2028.

The market-by-market clocks

Run the three clocks against the individual markets and a rough sequencing emerges.

Dallas-Fort Worth looks earliest. Its rate declines have moderated to under 1% year over year, its pipeline has normalized, and Extra Space already reports positive move-in rates in Dallas. The market’s problem was always more cyclical than structural, and a return to composite NOI growth during 2027 looks plausible.

Houston is the complicated one. Its pipeline is no longer outsized, which argues for cyclical repair on a similar schedule, but its per-capita inventory above 10 square feet is a structural fact that no absorption cycle erases. Houston can stop getting worse in 2027; the last cycle suggests its rates recover to a ceiling lower than its Sunbelt peers would accept, absent Texas-scale population growth doing the heavy lifting for another decade.

Tampa, Orlando and Atlanta sit in the middle. All three still show falling rates, but deliveries are decelerating fastest in exactly these metros, and analyses this year have singled them out as the markets where the supply slowdown is most visible. If 2026 marks their delivery peak-to-trough inflection, the three-year lease-up shadow points to pricing stabilization during 2027 and real rate growth in 2028.

Phoenix is later. At 6.9% of stock under construction, the market is still adding supply at triple the national pace, with 2.9 million square feet arriving this year alone, about 7% inventory growth in a single year. The pipeline says deliveries continue well into 2027, and the lease-up clock pushes rate pressure into 2028 and possibly beyond. Phoenix’s saving grace is demand: its population and job growth remain among the strongest of any major metro, which shortens lease-ups. But buyers underwriting a 2027 Phoenix recovery are underwriting hope.

Cape Coral-Fort Myers is last, and it is not close. The metro is adding roughly 12% to its inventory this year, on top of the post-hurricane building boom, with 5.4% of stock still under construction. An NOI decline of 14.3% is what a market looks like when supply growth runs at more than four times the national rate. CubeSmart chief executive Chris Marr called Fort Myers and Cape Coral “multi-year recovery projects” on his July earnings call, and the math backs him up: full stabilization before 2029 would require something extraordinary from Florida in-migration.

What could move the dates

Two developments could compress these timelines, and one could extend them.

The first accelerant is housing. Every date above assumes steady, unspectacular demand. A genuine thaw in existing-home sales, which no one forecasts before late 2027 at the earliest, would function exactly as the pandemic did in Texas: a demand wave landing on a shrinking pipeline. The overbuilt markets, which skew toward exactly the high-migration metros where housing turnover recovers first, would paradoxically benefit most.

The second is capitulation. As unstabilized 2023-24 vintage facilities change hands at discounts to development cost, the reset basis lets new owners stop discounting sooner than a seller carrying a full construction loan ever could, and each of those trades shortens the market’s repricing cycle a little.

The risk runs the other way in the pipeline data. Yardi Matrix raised its national supply forecasts this year for the first time since 2023, now projecting 51.1 million square feet in 2026, 44 million in 2027 and about 38 million in each of 2028 and 2029, with the 2029 figure revised up 15%. The firm still models supply growth near a historically low 1.7% of stock through 2031, but the floor appears to be in, and the metros where pipelines re-expanded this spring include Sunbelt names like San Antonio, Nashville and Las Vegas. If cheap land and falling construction backlogs tempt developers back into the recovery markets before they have recovered, the clocks reset.

The composite picture for the sector’s laggards, then, is not grim so much as staggered: Dallas in 2027, Tampa and Orlando behind it, Phoenix in 2028, Cape Coral at the end of the decade, and Houston cyclically sooner but structurally capped. The overbuilt Sunbelt will heal the way it always has, one lease-up at a time, unless the housing market shows up early and does it the fast way. For buyers, the staggering is the opportunity. The best storage deals of the last cycle were signed in overbuilt markets by investors who could read a delivery schedule, and the delivery schedules have rarely been this legible.