The Tunnel Boom Has Peaked. Here Is Exactly Where.

• 

  Vehicle passing through an express car wash tunnel conveyor

Industry revenue is still climbing. So is the number of stores. Underneath both, the business that new entrants actually depend on has been shrinking for four straight quarters.

On February 5, 2025, Zips Car Wash filed for Chapter 11 protection in the United States Bankruptcy Court for the Northern District of Texas. The numbers in the petition were not subtle. The company carried 653.9 million dollars of funded debt against roughly one million dollars of cash. It operated more than 260 locations across 23 states under the Zips, Jet Brite and Rocket Express brands, and it had roughly 625,000 people paying monthly for unlimited washes.

Twenty days later, Driven Brands announced it was leaving the business entirely. It sold its whole US car wash operation to Express Wash Operations, trading as Whistle Express, for 385 million dollars, comprising 255 million in cash and a 130 million dollar seller note. The deal closed on April 16. Chief executive Jonathan Fitzpatrick framed it as debt reduction and a refocus on the Take 5 Oil Change brand. By late 2025 Driven had agreed to sell its international car wash business too, for roughly 406 million euros, and had reclassified the entire segment as discontinued operations.

Two of the largest players in American car washing spent the same spring reducing their exposure to it. One through a courtroom, one through a sale.

The bankruptcy filings named the cause directly. Increased competition, driven by an estimated 900 new washes opening annually over the prior five years.

That is the story most coverage tells, and it is roughly correct. But it misses the more useful part, which is that the national statistics still look fine. Industry revenue including detailing runs near 18.7 billion dollars. Same-store sales at benchmarked locations are up. Chains are still opening stores. If you read only the headline numbers you would conclude the sector is healthy and these were two badly capitalized companies.

The headline numbers are hiding something specific, and once you see the mechanism you cannot look at a trade area the same way again.

The revenue that is growing is not the revenue a new site can win

Rinsed, a software provider whose quarterly benchmark reports cover roughly 3,000 to 3,500 locations, publishes a breakdown that almost nobody outside the industry reads. It separates membership revenue from retail revenue, meaning what walk-up customers pay.

The two lines have been moving in opposite directions for a year and a half.

In the second quarter of 2025, total same-store sales rose 5.6 percent. Membership revenue rose 15.2 percent. Retail revenue fell 11.9 percent year over year, and it fell despite a 3 percent price increase, which means volume dropped further than the revenue figure alone suggests. The third quarter showed same-store sales up 5.7 percent, membership up 10, retail down 3. The fourth quarter: sales up 4.8, membership up 11.3, retail down 8.4. The first quarter of 2026: sales up 5.8, membership up 10.6, retail down 3.3, and stores that had been open two years or more grew only 2.5 percent.

Four consecutive quarters where the top line rose and the walk-up business shrank.

Some of that is weather. Rinsed found the typical location saw roughly 10 percent more rainy days in the second quarter of 2025 than in the same quarter of 2024, which suppresses retail volume without touching a monthly subscriber. Do not overstate the structural read. But three more quarters followed the same pattern under different weather, and the operator data corroborates it. Mister Car Wash reported unlimited wash club sales running at 76 to 79 percent of total wash sales in late 2025 and early 2026, up from 73 to 75 percent a year earlier, with membership growing 7 to 11 percent while comparable store sales grew between 1.6 and 3.9 percent.

Now consider what that means for a developer standing on a vacant pad site with a tunnel in mind.

A new car wash cannot inherit a membership base. It opens with zero subscribers and has to convert them out of the walk-up traffic that is currently shrinking. And conversion is not uniform. Rinsed's data shows sites with more than 4,000 members converting roughly 11 to 15 percent of their retail customers into subscribers. Sites with fewer than 2,000 members convert between 2.4 and 3.4 percent.

Read that twice. The stores that already have members are four to five times better at getting more of them. Membership is a flywheel that rewards whoever spun it first, and a late entrant in a crowded corridor is trying to start one from a standstill using the half of the market that is contracting.

That is what saturation looks like from inside a growing income statement. The incumbents are fine. The next tunnel is not.

Nobody agrees how many car washes exist

Before you can measure saturation you have to count the denominator, and here the industry has a genuinely interesting problem.

The Census Bureau counted 17,584 car washes with paid employees in 2020. A separate Census tally put establishments at 18,520 the same year. Employer revenue for the category ran about 16.3 billion dollars in 2022.

The International Carwash Association, counting by machine format, arrived somewhere else entirely. Its 2020 study found roughly 17,500 conveyor tunnels, 29,000 in-bay automatics and 16,250 self-service wash operations. That sums to about 62,750 sites.

Same country, same year, and a difference of a factor of three and a half.

Neither is wrong. They are counting different things, and the three reasons they diverge are worth understanding because each one has consequences for a market study.

The first is employment status. Census establishment counts capture businesses with payroll. A great many self-service bays and some in-bay automatics have no employees at all, so they never appear. That single distinction explains most of the gap.

The second is where a wash bay lives. A wash attached to a fuel station or convenience store is frequently classified under the parent site, not under car washing, so it disappears into gasoline retailing.

The third is definitional. The association counts wash points by machine type. Census counts business establishments. One tunnel business is one establishment. A bank of six self-service bays might be one establishment, or six tax filers, depending on how it is structured.

For a saturation study, this forces a decision that most market analyses skip: what are you actually counting? If the question is whether another express tunnel can win members, the relevant universe is the conveyor count and the branded chain rosters, not the all-in 62,750. Self-service bays are not competing for the same monthly subscriber. Counting them makes a market look more crowded than it is for your format, and counting only employer establishments makes it look emptier.

The code survived, which is better news than it sounds

Here is a piece of plumbing that determines whether any of this can be measured at the county level.

When the North American Industry Classification System was revised for 2022, a lot of retail codes were renumbered and several were consolidated. Car washes were not. NAICS 811192, Car Washes, survived the revision intact as a distinct six-digit national industry. So did its sibling, 811191, covering automotive oil change and lubrication shops. Neither was folded into a broader catch-all.

That matters more than it sounds. It means a continuous, county-level series exists, and it means the federal government publishes establishment counts for exactly the business you are trying to site.

The code sits in Sector 81, Other Services, under subsector 811 for Repair and Maintenance and industry group 8111 for Automotive Repair. Its definition covers establishments primarily engaged in cleaning, washing and waxing automotive vehicles, and it sweeps in conveyor tunnels, in-bay automatics, self-service wand bays, hand washes, detailing, mobile operations and truck washes alike.

Four free federal datasets carry it, and each answers a different question.

The Bureau of Labor Statistics Quarterly Census of Employment and Wages is the workhorse. It publishes establishment counts at the six-digit code for every county in the country, drawn from state unemployment insurance records covering roughly 95 percent of American wage and salary jobs. It updates quarterly with about a five to six month lag. Two limitations matter. A large share of county-level cells are suppressed to protect confidentiality, and across all industries that suppression touches something near 60 percent of private county data. And because it comes from payroll records, it misses most non-employers. The workaround for the first problem is the QCEW Data Viewer's option to show records with suppressed employment and wages, which tells you a county has activity even when it withholds the count.

County Business Patterns publishes annually, employer establishments only, with something QCEW does not offer: establishment counts broken out by employment size class, which is a crude but real proxy for format. A one-to-four-employee establishment is probably not a 150-foot tunnel with twenty vacuum stalls. The most recent release covers 2023, with 2024 due in summer 2026. Its ZIP Code Business Patterns companion takes the same content down to the ZIP level, which matters because trade areas do not respect county lines.

Nonemployer Statistics fills the hole the other two leave. Built from IRS business income tax records, it publishes establishment counts and receipts by industry code down to the county, currently through 2023, with a series running back to 1997. This is where the self-service washes live. One caution: it also includes mobile detailers who operate no fixed site at all, so it overstates physical locations.

The Economic Census, every five years and currently on its 2022 vintage, adds revenue by establishment at county level.

Put together, those four give you an honest count of car wash businesses in any American county, for free, with a known and describable error band.

What none of them give you is format. No federal dataset codes whether a car wash is a tunnel, an in-bay or a self-service bay, and none identifies chain ownership. There is no published figure anywhere for express tunnels per county. If you want that number, you build it. The recipe is at the end of this article.

The build rate already turned

New express car wash development fell from 943 openings in 2022 to approximately 550 in 2025. Advisory projections for 2026 run to 450 or fewer, though that is a forecast rather than a count and should be treated as one.

Which brings up something worth correcting, because it appears constantly in trade coverage and in investor decks.

The figure of roughly 900 new washes a year is the 2021 to 2022 peak. It is not the current run rate, and it has not been for three years. The Zips bankruptcy filings cited it as the annual pace over the prior five years, which was accurate as a historical average of a period that included the boom. Repeated today as a present-tense statistic, it overstates new supply by roughly 60 percent. The association's own baseline material describes new construction as exceeding 500 units a year, which is closer.

The chains, meanwhile, are still opening stores, just fewer of them, and the ownership map has consolidated hard.

Mister Car Wash ended 2025 with 548 locations and stood at 549 as of March 31, 2026. It opened 16 greenfield sites and acquired five in the fourth quarter alone, netting roughly 47 additions across the year, and crossed a billion dollars of revenue for the first time. It is now being taken private by Leonard Green and Partners, and it canceled its February 2026 earnings call and withheld guidance because of the transaction.

Whistle Express, after absorbing the Driven Brands portfolio, operates roughly 530 locations across 23 states, backed by Oaktree. The combined Club Car Wash and Express Wash Concepts platform runs about 389 sites. Quick Quack operates somewhere between 275 and 350 depending on whether you count sites in development, took a large investment from KKR in 2024, and bought 13 Utah locations from Whistle in October 2025. Tidal Wave crossed 300 locations in October 2025 and stood at 303 by December. Tommy's Express runs roughly 256 to 270 on a franchise model. Below them sit Mammoth at around 250, Spotless Brands at 213, Zips at roughly 197 after restructuring, GO at 154, WhiteWater at 144, ModWash at 130, Super Star at about 118, El Car Wash at 90 to 100 in Florida, and Autobell at 91.

The character of dealmaking has changed with the pace. The most active buyers in 2025 and 2026 have been acquiring distressed single sites, small chains and entitled parcels below replacement cost. That is not a growth market. That is a market where the cheapest way to add a location is to buy someone else's mistake.

The demand ceiling nobody talks about

There is a structural reason saturation arrived when it did, and it has nothing to do with interest rates.

According to the International Carwash Association, the share of American drivers who wash their cars most often at a professional facility rose from 50.0 percent in 1996 to 79.0 percent in 2023.

For nearly thirty years the industry had a growth engine that required no new customers, only conversion. Every year, some percentage of people who used to wash the car in the driveway stopped doing that. The tunnel operators were not fighting each other for a fixed pool. They were splitting a pool that kept growing.

At 79 percent, that engine is close to spent. Four in five drivers are already converted. The remaining fifth includes people who cannot be converted at any price, because they lack a car, live in dense urban cores, or simply will not pay.

From here, growth in any given trade area comes almost entirely from taking someone else's customers. That is a different business, and it is the business every new tunnel is now entering.

The math on whether a site works

The industry uses two rules of thumb, and both are crude, and both are more useful than nothing.

The first counts vehicles per wash within the trade area. Above 4,000 registered vehicles per wash suggests an underserved market. Between 2,500 and 4,000 is balanced. Between 1,500 and 2,500 is competitive. Below 1,500 is oversaturated.

The second divides three-mile-radius population by the number of tunnels, including the proposed one. Trade guidance treats roughly 33,000 people per tunnel as the saturation line, which is to say a market of 100,000 supports about three well-run express washes and not a fourth.

Underwriting a competition-free ground-up express typically assumes a capture rate of only 0.5 to 1.0 percent of passing traffic, and that rate falls as competitors appear.

Vehicles per wash in trade area	Reading	What it means for a new tunnel
Above 4,000	Underserved	Proceed to trade-area analysis
2,500 to 4,000	Balanced	Proceed only with a hard corner and a clearly superior offer
1,500 to 2,500	Competitive	High risk, and only if you can be first with a differentiated format
Below 1,500	Oversaturated	Do not build. Look at acquisition instead

Against those ratios sits the volume the building actually needs. Appraisal practice puts breakeven near 112 to 125 cars per day for every million dollars invested. A 2.5 to 3 million dollar express therefore needs something over 300 cars a day, every day, all year, before it earns anything. Annual conveyor volumes commonly center near 75,000 washes with a standard deviation around 25,000, which tells you how wide the outcomes are.

The clearest field signal that a corridor has tipped costs nothing to observe. When three, five and seven dollar washes appear in a market where the cheapest wash used to hold near ten, capacity has outrun demand and operators are buying volume. Drive the corridor and read the signs. It is a better indicator than any dataset because it is what the competitors themselves believe.

The permission problem arrived on schedule

Municipalities noticed the build-out before most investors did, and the ordinance response has been fast and geographically broad.

Hemet, California went furthest. In August 2024 it adopted an ordinance discontinuing new full-service and self-service car washes in any commercial or manufacturing zone, a permanent citywide prohibition. Hesperia, in the same county, approved a 45-day moratorium in February 2024 extendable to about ten months, after counting 19 existing washes plus five more approved since 2019. A council member asked staff for a saturation map, which is a notable thing for a city to want.

Birmingham, Alabama enacted a moratorium in March 2024 and extended it through 2025, responding to a surge of automated wash applications. Cape Coral, Florida followed in April 2024 after finding 16 existing washes and 12 more in the pipeline, and considered both one-mile minimum spacing and per-capita caps. Greenacres, Florida simply doubled its required separation between wash facilities, from 1,500 feet to 3,000. Warren, Michigan paused new car washes and gas stations together pending zoning amendments.

Elsewhere the change is quieter and more consequential. Orange Park, Florida and Hanover County, Virginia converted car washes from by-right uses to conditional uses, which does not prohibit anything but moves every application in front of a board and adds months. Fort Worth requires a conditional use permit within 200 feet of residential, prohibits dryers facing residential districts, and bars drying and vacuum equipment within 25 feet of any residential district. Los Angeles prohibits car washes within 500 feet of residential zones under its Clean Up Green Up overlay. Mayfield Heights, Ohio sets a 150-foot minimum lot frontage, a 300-foot minimum depth, and eight stacking spaces per tunnel entrance, which quietly disqualifies most infill parcels.

The legal risk runs both directions. A Rio Grande Valley municipality adopted a six-month moratorium in March 2024 and drew a lawsuit under the Texas vested-rights statute. A 2025 state law requiring two public hearings 30 days apart before adopting a development moratorium prompted the city to consider rescinding.

The pattern across all of it is the same. By-right approval is being replaced by conditional use, distance separation and per-capita caps, and the trigger is community perception of saturation. Water reclamation requirements in California, Nevada, Florida and Arizona add a further layer, which our earlier feasibility guide covers in detail.

The parcels left over are the ones that cannot work

This is the part that compounds the problem, and it is the part a site planner sees first.

An express exterior tunnel with a 100 to 150 foot conveyor needs roughly 0.8 to 1.3 acres of rectangular land, and critically, one dimension of at least 225 feet. The building runs 4,000 to 6,000 square feet. Add 12 to 20 vacuum stalls at about 75 square feet each. Tommy Car Wash Systems specifies a 0.8 to 1.0 acre minimum. Harvey Miller, a past president of the International Carwash Association, has called one acre ideal.

Stacking is where sites fail. An express tunnel wants 12 to 15 or more vehicles of queue capacity, and each vehicle occupies about 20 to 22 linear feet of lane. That is 240 to 330 feet of queue, on a site that also has to accommodate two-way drive aisles at 24 to 26 feet, one-way aisles at 12 to 16, a pay lane, the tunnel itself and the vacuum court. New York City requires ten automobiles of stacking per washing lane in manufacturing districts. Once everything is paved, impervious coverage typically lands between 75 and 85 percent, which leaves very little room for the landscaping, buffering and stormwater that the conditional use permit will demand.

Corner lots with two access points outperform interior parcels consistently, because they solve ingress, egress and queue spillover at the same time. Traffic thresholds bear that out. A working express in Bixby, Oklahoma sits on a corridor carrying about 37,000 vehicles a day and washes roughly 109,000 cars a year. A Baton Rouge express operates on a four-lane corridor above 28,000 vehicles a day.

Here is the compounding effect. In a corridor that has already been built out, the sites still available are, almost by definition, the ones the earlier entrants passed over. Interior lots. Single access. Dimensions under 225 feet. Awkward frontage. Which means the late entrant is not merely competing against established membership bases. It is doing so from a physically inferior site that cannot achieve the throughput the incumbents achieve.

A compromised site in an unsaturated market can work. A good site in a saturated market can sometimes work. A compromised site in a saturated market is how you end up as somebody's distressed acquisition.

What the capital markets are saying

The net lease market has been sending the same signal through pricing.

Average car wash cap rates closed 2025 near 6.26 percent, down 38 basis points year over year, and stood at roughly 6.24 percent in January 2026 across about 200 listings with an average remaining lease term of 18.5 years. The average listed property carried a price near 4.99 million dollars and net operating income around 312,822 dollars.

Falling cap rates normally signal confidence. This time they arrived alongside a 36 percent quarter-over-quarter jump in listed inventory, driven largely by the reinstatement of 100 percent bonus depreciation in July 2025, which made car wash structures attractive to tax-motivated buyers as 15-year property. Demand for the paper went up. That does not mean demand for the underlying business did.

Look instead at where trades actually clear. Assets that printed in the mid-5 percent range at the 2022 peak now clear closer to 6.2 to 6.4 percent, roughly 100 basis points wider. Strong-credit Quick Quack properties have gone as low as 5.39 to 5.50 percent. Weaker tenants price above 7. The spread between the best and worst operator credit has widened, which is what a market does when it starts distinguishing between businesses instead of buying a category.

For individual buyers, Small Business Administration lending remains the dominant path, with 7(a) and 504 programs and lenders including Live Oak Bank, Ready Capital and Celtic Bank most active. The 504 structure generally produces the lowest revenue breakeven; 7(a) offers more flexibility at a higher hurdle. Either way, the debt service assumption should be tested against the 300-cars-a-day threshold, not against a proforma built on a first-mover's volumes.

A screen you can run before you option the land

Four stages, in order, and the first two cost nothing but time.

Stage one, the county screen. Pull QCEW annual establishment counts for NAICS 811192 by county. Add Nonemployer Statistics establishments for the same code to capture the self-service and no-payroll sites. Divide county vehicle registrations, available free from Federal Highway Administration Highway Statistics and state motor vehicle departments, by the combined count. Any county below roughly 2,000 to 2,500 vehicles per wash is presumptively oversupplied. Above 4,000 and growing, keep going.

Stage two, the trade area. County averages hide corridor clustering, and clustering is the whole problem. Re-run the ratio on a three-mile drive-time polygon using ZIP Code Business Patterns, and build a tunnel roster by hand, because federal data will not tell you which of those establishments are express conveyors. If your proposed tunnel would be the third or later within one mile, treat it as high risk regardless of what the county number says. If destructive pricing is already visible in the corridor, stop.

Stage three, the geometry gate. Require 0.8 to 1.3 acres, a minimum dimension of 225 feet, a hard corner or dual access, and traffic counts consistent with the latent demand your screen implies. Pull average annual daily traffic free from the Federal Highway Administration's Travel Monitoring Analysis System or from state portals including Florida Traffic Online, Washington's data , Maryland's at data and Utah's. Reject interior, single-access or undersized parcels even in unsaturated counties. A site that cannot stack 12 vehicles cannot hit 300 cars a day on a Saturday, and Saturday is the whole business.

Stage four, what to do when the screen says no. Saturated counties are not dead markets, they are markets where new construction is the wrong instrument. Buy distressed or under-membered existing tunnels below replacement cost and fix operations, which is what the most active acquirers are already doing. Build mini-express or in-bay formats on the 0.25 to 0.5 acre infill parcels that a full tunnel cannot use. Redeploy toward secondary metros and snow-belt corridors that the 2020 to 2025 wave under-served. Or repurpose an entitled car wash parcel toward an adjacent convenience use, since the entitlement itself often carries more value than the wash it was obtained for.

What would change this call

Three things, and they are worth watching because none has happened yet.

If benchmarked retail revenue turns positive year over year for two consecutive quarters, the walk-up business is recovering and the flywheel argument weakens. If net express openings re-accelerate above roughly 700 a year, operators are seeing something the current data does not show. If net lease cap rates compress below 6 percent while listed inventory falls, capital is genuinely re-rating the asset class rather than chasing a depreciation schedule.

As of the first quarter of 2026, retail revenue is still negative, builds are still decelerating, and the inventory overhang persists.

The tunnel boom was real, and it was rational while four in five drivers were still converting. What it produced is a country where the average corridor now has enough capacity and the average available parcel is the one nobody wanted. Both of those are fixable problems for an operator with a membership base. Neither is a fixable problem for a new build.

Run the screen before you option the land. It costs a week and a few federal downloads, and it is considerably cheaper than finding out from a lender.

Sources:

• Federal dataUS Census Bureau County Business Patterns and ZIP Code Business Patterns

• Nonemployer Statistics and the Economic Census, all under NAICS 811192

• Bureau of Labor Statistics Quarterly Census of Employment and Wages

• Federal Highway Administration Highway Statistics and state department of transportation traffic count portals

• Company and court recordsMister Car Wash quarterly and annual disclosures.

• Driven Brands transaction announcements

• Zips Car Wash Chapter 11 filings, United States Bankruptcy Court for the Northern District of Texas, February 2025.