Express Tunnel Car Wash in Douglas County, Colorado: An SBA 7(a) Model Study

Model study: an illustrative feasibility analysis built on public data for Douglas County, Colorado; not a client engagement.

Project summary

All figures in this table are model results.

The credit question

A lender is not mainly asking whether people in Douglas County wash their cars; the county's income and vehicle counts make that the easy part of the case. It is asking four narrower questions, and each one shapes the structure.

First, can a single new tunnel reach a membership base large enough to cover debt service within two years of opening, when SBA's rule for start-ups requires projections showing debt service coverage of at least 1.15 within 2 years from the end of construction [1]? Membership is a stock that builds through sign-ups and drains through cancellations, so the ramp depends as much on churn as on marketing.

Second, does the project fit inside the 7(a) program at all? The maximum 7(a) loan is $5,000,000 and SBA's guaranty to one borrower and its affiliates cannot exceed $3,750,000 [2]. In this model study's budget, a new express tunnel on a Douglas County arterial pad costs $8.4 million all in. A 7(a) loan can therefore fund only about 59% of this project, and the rest must come from equity or from a second loan that is structured so SBA does not end up in a junior position.

Third, what is the collateral worth if the business fails? SBA lists car wash businesses among its examples of limited or special purpose property [1]. The building is a tunnel with pits, trench drains, reclaim tanks and a conveyor. Its alternative use is poor, and the appraisal a 7(a) lender relies on may not include the value of intangible assets [1]. The membership base, which is what makes the business valuable, does not count as collateral.

Fourth, is the market already full, and will it stay open to this operator? Tunnels compete for the same membership households, and at least one Douglas County city moved in 2025 to pause new car wash approvals [3]. Front Range water providers also restrict use in droughts. A lender needs a competitor method that it can audit and a water plan that survives a drought stage.

The short answer: the modelled project works at 30% equity and fails the SBA projection test at the 10% minimum.

Location and county economy

Population and growth

Douglas County sits between Denver and Colorado Springs along Interstate 25. Its population has grown every year this decade. Census Bureau estimates put the July 2025 population at 399,396, up from 360,315 in 2020 [4] and 322,333 in 2015 [5].

Source: U.S. Census Bureau population estimates, Vintage 2025 for 2020 to 2025 [4] and the 2010 to 2020 series for 2015 [5]; growth rates computed.

Growth is still positive, roughly 6,000 to 9,500 new residents a year, but it slowed to 1.61% in 2025. A model that carried the 2.6% pace of 2021 into member growth would overstate demand, so this study holds membership flat after stabilization and lets revenue grow only through price.

Households, income and vehicles

The membership model runs on households and vehicles, not on population. The 2024 American Community Survey (1-year) counts 147,825 households in the county, with an average household size of 2.65, a median household income of $151,659 and per capita income of $73,061 [6]. Owner-occupied homes have a median value of $742,700, and 4.8% of residents had incomes below the poverty level [6].

A membership plan is a fixed monthly card charge that pays off for households washing more than about twice a month. It sells best to owner households with steady incomes and several vehicles. Just over half of Douglas County households earn $150,000 or more, nearly four in five own their homes, and 27.7% have three or more vehicles, which makes discounted family add-on plans a real revenue line. These facts support a firm price. They do not, by themselves, support a higher member count, because competitors read the same census tables.

The work-from-home share, 29.6% in 2024 [6], cuts two ways: fewer commute miles mean less road grime, but home-based workers can wash mid-day, which spreads demand away from weekend peaks.

With an average household size of 2.65 [6], 6,000 to 8,500 new residents a year imply roughly 2,300 to 3,200 new households and, at 2.08 vehicles each, about 4,700 to 6,700 more vehicles a year in the county.

Employment

Douglas County is mainly a place where people live. Resident employment by industry (ACS 2024 1-year) shows a white-collar base [6].

Source: [6]; shares computed.

A car wash membership is a small discretionary subscription, and cancellations rise when budgets tighten. That argues against an aggressive price path, so the model raises prices 3% a year, an illustrative assumption, and treats higher churn as a core stress test. The spread of professional and health care employment limits the risk that one employer's layoffs move churn across a trade area.

Commuting and traffic

Retail (single) washes come mostly from passing traffic, and members are recruited from drivers who pass the site on routine trips. Colorado DOT publishes annual average daily traffic (AADT) for state highway segments through its Online Transportation Information System, where each station also carries a 20-year factor [7]. The 2025 AADT estimates for Douglas County segments show where volume sits.

Source: CDOT OTIS traffic stations, Douglas County, AADT year 2025 [7].

The arterials serving the county's main towns carry 24,000 to 48,000 vehicles a day. A car wash needs access on one of these; I-25's 139,000 vehicles pass at speed and add visibility, not customers. Several values are factored from older counts (SH 83 south of Hilltop was last counted in 2019), so a lender should ask for a current turning-movement count at the access point. CDOT's 20-year factors of 1.09 to 1.23, read as growth over 20 years, imply traffic growth of about 0.4% to 1.0% a year, slower than recent population growth: traffic will not fill the tunnel, membership will. This study's site screen, an assumption, is frontage AADT of at least 25,000 on the going-home side of an arterial, with signalized or full-movement access.

Market analysis

Demand drivers

Climate and road treatment. The NOAA 1991 to 2020 normals for the Castle Rock station (USC00051401) record 57.5 inches of snow a year, an average of 17.0 days with at least 1.0 inch of snowfall, 6.7 days with at least 3.0 inches, and 166.7 days a year with a minimum temperature below freezing [8]. The Highlands Ranch 3.2 ESE station (US1CODG0074), at the north end of the county, records 64.0 inches and 18.7 days with at least 1.0 inch [8]. Spring is a heavy snow season: March and April together average 19.0 inches at Castle Rock [9].

Climate columns: NOAA NCEI normals, Castle Rock (USC00051401) [8], [9]. The demand index and wash columns are model results: the index is 1 + 0.08 × snow days, normalized to 12.

Colorado DOT has described treating state highways with liquid magnesium chloride, a cold-temperature modified magnesium chloride for colder pavement, and solid granular salts mixed with magnesium chloride such as Ice Slicer. Its release describes magnesium chloride as effective down to about 16°F of pavement temperature, the modified version below approximately 15°F, and Ice Slicer mainly at 20°F and below [10]. That 2012 release covers another CDOT region and is used only for the chemistry, not for quantities or current practice in Douglas County. Brine films on lower panels and undercarriages create a reason to wash after each storm, and an underbody flush is a natural upsell for top tiers.

The credit point is timing more than volume. Storms create surges on the first clear day afterwards. Members absorb those surges inside a flat fee, while retail washes swing with the weather, so a mild, dry winter appears among the downside scenarios. The long frost season also raises cost: tunnel heating, heated slabs and more dryer time.

Membership economics. No public source reports membership counts or revenue for Front Range tunnels, so the membership ramp, churn and price below are labelled model assumptions. A single-site independent should not be assumed to beat established operators without site evidence, and the model treats its steady-state member count as an upper bound.

Existing supply and pipeline: method plus assumption

No public registry lists conveyor car washes by county with tunnel length, so supply is built by method:

1. Define the trade area as a 3-mile ring, adjusted to a 10-minute drive time, around the site.
2. Inventory every express or flex tunnel of 80 feet or more by field visit and operator websites, recording tunnel length, plan prices and opening year. In-bay automatics are counted separately; they compete for retail washes, not members.
3. Search planning-commission and council agendas in each jurisdiction that touches the ring (Douglas County, Castle Rock, Parker, Lone Tree, Castle Pines and the Highlands Ranch area) for approved or pending car wash site plans in the past 24 months.
4. Count the subject plus each competitor tunnel and compute vehicles per tunnel and the subject's required share of the membership pool.

The pipeline step matters here. The City of Castle Pines gave public notice that its council would consider, on March 25, 2025, an ordinance establishing a temporary moratorium on approval of land use applications and business license applications for car washes and several other uses in all zoning districts of the city [3]. A moratorium slows competition inside that city, but it also shows that a council can stop a car wash entitlement at short notice, so the loan should not close before site plan approval is final and appeal periods have run.

This model study assumes three existing express tunnels in the trade area plus one more opening within 24 months of the subject. That count is the first item a site-specific study would replace.

Demand method: membership pool and fair share

The method is a fair-share test on the membership pool; retail washes are modelled separately.

The subject needs 84% of its fair share today and 105% once the pipeline tunnel opens. An index near 1.0 is defensible for a new tunnel with good access, but not if its access or visibility is weaker than its competitors'. The test is sensitive to plan penetration: at 18% of vehicles on a plan, the five-tunnel index is 1.17; at 22% it is 0.96. Until site data replace these inputs, this study treats 4,211 members as an upper bound and tests lower counts under the stress tests.

Membership ramp and churn

Members are modelled monthly as a stock: starting members, less cancellations (a percentage of the starting stock), plus new sign-ups. Steady state equals sign-ups divided by the churn rate.

Year 1 churn is higher because launch offers (a discounted first month) pull in members who cancel when the full price starts. At 5.70% monthly churn and the Year 3 average price of $29.71, a member is worth about $521 in gross revenue over an average 17.5-month life. One extra point of monthly churn cuts that life to about 15 months, which is why churn is the variable to watch.

The model reaches 3,056 members at the end of Year 1 and 3,781 at the end of Year 2, then flattens toward 4,211. The ramp assumes new households replace cancelling ones, not that membership grows with the county. The price is the blended realized price across tiers and family add-ons; the $28.00 starting figure, set for a 2028 opening, is an assumption to check against competitor price boards.

Forecast

The forecast combines the membership stock, retail washes of 30,000 in Year 1 (opening traffic), 28,000 in Year 2 and 27,500 in Year 3, growing 1% a year thereafter, and ancillary sales of 2% of wash revenue. Members wash 3.0 times a month; that drives cost and throughput, not revenue.

At Year 3, the model processes 167,966 vehicles a year, or 460 a day on average. The peak month (February) averages 557 a day. On a clear day after a storm, demand can reach about twice the monthly average, about 1,114 cars, with 12% of those in the busiest hour, about 134 cars. A 130-ft tunnel handles about 110 to 125 cars an hour (assumption), so the site will queue for an hour or two on its best days. The site plan therefore needs at least 25 stacking spaces and a member lane with licence-plate recognition, and the longer tunnel suits this market better than a 100-ft one.

Site and technical

Site requirements

The modelled site is a 1.30-acre pad. The criteria a lender should test are:

• Frontage AADT of 25,000 or more on a Douglas County arterial, with the site on the going-home side.
• Full-movement or signalized access, with right-in, right-out as the minimum.
• Lot depth of about 330 feet or more, so queues never reach the public street.
• Central water and sewer at the lot line, with capacity confirmed in writing by the provider.
• No prior site use that would push the environmental review past a Transaction Screen (see program compliance).

Zoning and permits

Unincorporated areas follow the Douglas County Zoning Resolution; Castle Rock, Parker, Lone Tree and Castle Pines run their own codes. A site-specific study must confirm the parcel's zoning district, whether a car wash is a use by right or needs a written use determination, any design rules, whether central water and sewer service is required, and whether a moratorium such as the Castle Pines ordinance [3] applies. A written use determination should come before design spending.

Inside municipalities the usual path is site or development plan review with a hearing, then building, grading and stormwater permits.

Water rules and reclaim in Colorado

Water is the Colorado-specific issue a lender should not take on trust. Drought stages and any limits on commercial car washes are set by each local water provider, so the file needs the provider's current drought plan for the actual site. Two published standards help with design:

• Colorado Springs Utilities, a Front Range utility, runs a car wash certification program under which conveyor and express washes must have a recycling system, limit water use to 35 gallons per vehicle on the base wash cycle, and recycle any spot-free reverse osmosis reject water [11]. It does not bind Douglas County but is a sound design target.
• Colorado approves reclaimed water from centralized systems for vehicle washing, automated and manual, as a Category 2 use, which requires secondary treatment with filtration and disinfection, according to EPA's summary of the state regulation [12]. That matters only if the provider delivers reclaimed water; on-site recycling is a separate matter.

The modelled design is a $165,000 reclaim system that recycles wash water, rinse water and RO reject, using fresh water for the final and spot-free rinse, with a target of 35 gallons of fresh water per car: about 5.9 million gallons a year at 167,966 cars. The Year 1 water and sewer cost of $0.60 per car implies about $17 per 1,000 gallons, an assumption to replace with the provider's tariff. Drought enters the model through cost (the downside case raises water cost by half) and could, in a severe stage, restrict washing. Metered use per car and a working recycling system are the best mitigation, and the loan agreement should require them.

Construction approach and schedule

The building is a masonry or pre-engineered tunnel with an equipment room, reclaim vault, trench drains to a sand and oil interceptor and heated entry and exit slabs; equipment comes as one supplier package. Construction runs 12 months from closing under a fixed-price contract. For a 7(a) construction component of more than $350,000, the SOP requires evidence of a 100% performance bond and a labor and materials payment bond, with builder's risk and workers' compensation coverage, unless the lender retains a third-party construction management firm with funds control or has its own construction management department [1].

A late-summer opening captures storm-season sign-ups in months 2 through 8; a spring opening would meet weaker summer demand first.

Development budget and sources and uses

All budget lines are assumptions (see the assumptions register) except the SBA guaranty fee, which applies the SOP's maximum fee chart to a $3,750,000 guaranteed portion [1]. Construction-period interest assumes equity is spent first and the two loans are then drawn evenly over 12 months at their note rates; it is solved iteratively because the interest is itself financed.

Real-estate uses (land, site work, building, taps, design, permits, monitoring and their share of contingency) total $5,156,703, or 61.3% of project cost. Equipment uses with their contingency share total $2,085,500 (24.8%), and the remainder is soft, financing and working capital (14.0%). The real-estate share matters for maturity, discussed under debt structure.

Revenue build and operating expenses

Year-by-year revenue

Stabilization is Year 3, the first full year after SBA's two-year window. Revenue of $1.92 million in Year 3, with members producing 73.7% of wash revenue, is a model result for a single new site.

Operating expense structure

Fixed costs dominate: labour, property tax, insurance, software, snow removal and base energy do not fall with volume, so NOI swings more than revenue. Year 1 revenue of $1.24 million against operating costs of $965,864 gives a 21.8% margin; Year 3 reaches 38.6%. That gearing is why the break-even tests sit close to the base case.

Three lines need local confirmation: property tax ($100,000 in Year 1 dollars), which depends on the assessor's valuation and the parcel's mill levy; labour, at an hourly wage of $19.50 before burden; and water and sewer, which follow the provider's tariff. Management is imputed at 3.5% even for an owner-operator, because a lender should charge the cost of replacing the operator. Marketing is $120,000 and $70,000 in the launch years, then 2.5% of revenue. Repairs start at 1.5% of revenue under warranty and reach 4.0% by Year 4. The replacement reserve starts at $40,000 in Year 3 dollars. All expenses rise 3% a year.

Assumptions register

Every value in this table is a model assumption.

10-year pro forma

NOI is stated after the replacement reserve. Figures are pre-tax and before owner distributions. Year 1 cash flow after debt service is negative by $331,996, which the $450,000 operating reserve covers with $118,004 left at the end of Year 1. Cumulative cash flow turns positive during Year 4.

Debt structure under the program

SBA 7(a) loan: $5,000,000. That is the program maximum for a single 7(a) loan [2]. At a 75% guaranty for loans over $150,000 [1], the guaranteed portion is $3,750,000, which is also the aggregate SBA guaranty limit for one borrower and its affiliates [2]. Further SBA capacity would have to come mainly through 504, and the SOP says the 7(a) loan should be processed and approved first when the two are combined [1].

Rate. The bank prime rate was 7.00% on each business day from November 3 to 9, 2026 [13]. The model uses Prime + 2.50% = 9.50%, variable. SBA's maximum for variable-rate 7(a) loans of more than $350,000 is three percentage points over the base rate [2], so 10.00% at today's prime. SBA publishes the maximum allowable fixed rate in the Federal Register [1]; this study models the variable rate and tests increases under the stress tests.

Maturity. 7(a) real estate loans may run up to 25 years, and an additional period reasonably necessary for construction may be added [1]. For mixed-purpose loans, the maturity may be up to 25 years if 51% or more of the use of the 7(a) proceeds is for real estate [1]. Real-estate uses are 61.3% of project cost, and because the SOP requires the lender to verify the equity injection before disbursing any loan proceeds [1], the lender can document the 7(a) proceeds as funding real estate and site work. The model uses 12 months interest-only during construction followed by 300 monthly payments of $43,685.

Fees. The SOP's maximum upfront guaranty fee for loans of $700,001 to $5,000,000 is 3.5% of the guaranteed portion up to $1,000,000 plus 3.75% of the guaranteed portion above that [1]. On a $3,750,000 guaranteed portion that is $35,000 plus $103,125, or $138,125, financed in the budget. SBA may announce temporary fee changes for a specific fiscal year through an Information Notice [1], so the fee must be confirmed for the approval date.

Conventional loan: $892,365, pari passu. The SOP prohibits a "piggyback" structure, in which lenders make more than one loan to a borrower at about the same time for a similar purpose and the SBA loan is secured with a junior lien position or no lien [1]. It says a shared, pari passu lien is not a piggyback when the non-SBA loan's maturity is not shorter than the SBA loan's [1]. The conventional piece is therefore modelled with matched collateral, a matched lien position and a matched 25-year maturity, at 7.25% fixed with payments of $6,450 a month. A lender unwilling to hold that piece on those terms has two choices: lend only the 7(a) and require about 40% equity, or put its own piece behind the 7(a) in lien priority, since the piggyback definition covers only cases in which the SBA loan holds the junior lien or no lien [1].

Equity injection and its basis under the rules. The SOP treats a business as a start-up if it has been in operation, generating revenue from intended operations, for 1 year or less, and considers an equity injection of at least 10 percent of total project costs, meaning all costs required to become operational, necessary for a start-up [1]. On this budget, that floor is $841,766. The SOP also requires a determination that the equity position and pro forma debt-to-worth are acceptable given the type of business, management experience and the level of competition [1]. The 504 program offers a cross-check: 13 CFR 120.910 requires at least 15% from a borrower that has operated two years or less or whose project involves a limited or single purpose building, and at least 20% when both apply [2]. SOP 50 10 8.1 lists car wash businesses among its examples of limited or special purpose property [1]. That 20% rule does not apply to 7(a), but it shows how SBA prices special-purpose and start-up risk, and a 7(a) lender can reasonably use it as a reference. This model uses 30% because the SOP projection test, not a percentage rule, requires it. The table holds the operating case constant.

Project cost moves slightly because construction interest depends on the amount borrowed. At 25% equity the month-24 run rate is 1.14, just under 1.15; at 30% it is 1.21. That is the practical equity effect of a special-purpose start-up: the rules allow 10%, but the SOP's own projection test does not.

Guaranties, insurance and collateral. Any individual owning 20% or more of the applicant must give an unlimited full guaranty [1]. For collateral, the SOP values improved real estate at no more than 85% of market value when deciding whether a loan is fully secured [1]; the model values new equipment at 75% of cost, an assumption. If a loan is not fully secured, life insurance is required in the amount of the collateral shortfall for the principals of businesses dependent on one owner [1].

The appraisal for new construction must estimate market value upon completion [1] and may not include the value of intangible assets [1]. Special-purpose real estate can appraise below cost, because a buyer pays for a working business, not an empty tunnel. At 80% to 90% of real-estate cost, the modelled project is not fully secured on SBA's basis, so the lender should expect liens on guarantor real estate where available and life insurance on the key principal for the shortfall.

DSCR and break-even

The reserve is drawn for shortfalls and refilled before any distribution. The month-24 run rate annualises the business at the end of Year 2 (year-end members at Year 2 prices plus Year 2 retail); it is the figure to compare with the SOP's 1.15 test, which applies within 2 years from the end of construction [1]. The base case passes at 1.21. Year 3 debt yield is 12.6%.

Break-even tests (Year 3).

The margin is narrow: Year 3 volume can fall about 10% before coverage reaches 1.00 and only about 4% before it drops below 1.15. The operating break-even is wide, so the business would keep running at half its planned volume; the debt is the thin layer. That is why structure, not market, decides financeability.

Returns

The levered return comes mostly from the exit. Ten years of cash flow after debt service total $2,098,306, about 83% of the equity; the rest returns only on a sale or refinance. A lender should read that as a reason to restrict distributions, not to doubt the project.

The value of the stabilized business at Year 3 NOI and a 7.75% cap is $9,579,168, which puts the combined senior debt at 61.5% of going-concern value. The gap between that figure and the SBA collateral value above is the membership base: real to a buyer, absent from the collateral file.

Sensitivity and stress tests

Two-way grid: Year 3 DSCR by 7(a) rate and volume

Volume scales members and retail together, with variable costs; the conventional loan is fixed. Volume matters more than rate: a 10% shortfall costs about 0.23 of coverage, a 100 basis point rise about 0.08. A rate cap or partial fixed rate is still worth pricing.

Downside scenarios

A slow ramp is, in this study's judgment, the most likely failure and the costliest to carry: it needs $888,211 of support at its low point, about $438,000 more than the reserve. A new competitor does the most long-run damage, holding coverage near 1.0 through Year 5. Mild winters and drought pricing are smaller risks; each leaves Year 3 coverage above 1.0. The combined case fails and sizes the backstop a sponsor would need.

A 10% construction overrun does not change DSCR if the sponsor funds it, and the SOP requires the lender to obtain evidence of the borrower's ability to pay cost overruns before approving any contract modification [1]. Sponsor liquidity should cover an overrun and the slow-ramp shortfall beyond the reserve.

Program compliance

SOP 50 10 8.1 took effect on October 1, 2026, and applies to applications issued an SBA loan number on or after that date; SOP 50 10 8.0 continues to govern 7(a) and 504 applications submitted through September 30, 2026 [14]. The checklist below reads the modelled project against the current version.

Risks and mitigants

Model determination

The modelled project is financeable under SBA 7(a), but only with more equity than the program floor and with conditions that address the start-up ramp.

The market supports the asset. Douglas County has 147,825 households with a median income of $151,659, 2.08 vehicles per household and owner occupancy near 78% [6]. It added 39,081 residents in five years [4]. Its arterial corridors carry 24,000 to 48,000 vehicles a day [7], and the Castle Rock station averages 57.5 inches of snow and 17 days a year with an inch or more [8]. Each fact favours a membership car wash, but none proves the modelled membership of about 4,200, which rests on assumptions a site study must replace.

The structure is the constraint. The $5,000,000 cap on a 7(a) loan covers only 59.4% of an $8.4 million new build. At the 10% start-up minimum, the project would need a $2.6 million second loan and would show a Year 3 DSCR of 0.99, failing SBA's own 1.15 projection test. At 30% equity and a matched-maturity pari passu loan, the month-24 run rate reaches 1.21 and Year 3 reaches 1.23. The special-purpose listing does not raise the 7(a) equity floor directly, but it leaves the collateral short of full security on SBA's valuation basis, which pushes a lender toward more equity, guarantor collateral and life insurance.

A lender reviewing a file like this would likely set these conditions:

1. Cash equity of at least 30% of total project cost, verified and spent before the first loan draw.
2. A funded operating and debt service reserve of at least $450,000, held by the lender and released only after two consecutive quarters at a trailing DSCR of 1.25 or more.
3. Sponsor post-closing liquidity of at least $450,000 beyond the reserve, which covers the slow-ramp shortfall.
4. Final, non-appealable site plan approval and a written use determination before closing; written capacity and tap fee letters from the water and sewer provider.
5. A site-specific competitor inventory and pipeline search, and a trade-area household count, that support a required fair-share index of no more than about 1.0 with the pipeline included.
6. A reclaim system specified to a fresh-water target of 35 gallons or less per vehicle, metered, with a written drought operating plan.
7. A conventional companion loan documented as pari passu with a maturity not shorter than the 7(a), or a junior position, so the structure is not a piggyback.
8. Life insurance on the key principal for the collateral shortfall, unlimited guaranties from every 20% owner, and liens on available guarantor real estate.
9. Distributions restricted until the reserve is full and trailing DSCR has held at 1.25 or more for four quarters.
10. Evidence of express tunnel operating experience in the management team, or a management agreement with an experienced operator.

With those conditions the credit rests on what can be proven before closing: site, entitlements, water and sponsor capital. The membership ramp risk is then carried by the reserve and the sponsor, not the guaranty.