Travel Centers: AADT and Fuel Demand Models

Type truck stop demand analysis into a search engine and the results are services and studies: feasibility consultants, a state research report from 2003, a hydrogen-corridor optimization paper. What none of them hands the reader is the measurement layer underneath, the federal series that count the trucks, the parking spaces, the stations, the freight and the fuel price, each published on a schedule and citable to the row. This article assembles that layer for the travel center, the interstate-scale fuel and services property whose demand analysis is usually asserted with a traffic count and finished with folklore. The public record does better: it counts 195.8 billion combination-truck miles a year, surveys 11,696 drivers about where they can and cannot park, prices diesel weekly through a spring in which the price rose half again, and censuses the station industry down to the employment size of every operator. The models a lender needs are arithmetic on top of those series, and this article shows which series, which arithmetic, and where the public record genuinely ends.

A note on scope. Travel center here means the highway-oriented fuel-and-services format serving both cars and combination trucks; the corner gas station shares its NAICS code and appears in the same counts, and the analysis separates them by what separates them in fact, the truck-classified traffic and the parking field. This article is about measuring demand; it is not a feasibility study and does not price a site, a fuel margin or a revenue line.

The asset, counted: 110,000 stations, two codes, one recode

County Business Patterns for 2023 counts 96,002 establishments of gasoline stations with convenience stores, employing 848,306 people, and a further 13,172 other gasoline stations, the code that carries truck stops without convenience formats as well as standalone fuel retailers, with 168,654 employees (U.S. Census Bureau, County Business Patterns 2023, 2025). The convenience format is small-establishment retail, 35.5% of establishments under five employees; the other-gasoline code runs larger per site, reflecting the service formats inside it. The employment-insurance census tells the recent direction: gasoline stations with convenience stores grew from 99,522 reporting units on annual average in 2022 to 101,351 in 2024, employment from 862,925 to 883,550, while other gasoline stations slipped from 9,646 units to 9,277 (BLS, QCEW annual averages, NAICS 457110 and 457120, 2023 and 2025). The narrative of the disappearing gas station is not in the federal record; the convenience-format station count is growing at about 1% a year, and what shrinks is the fuel-only format the convenience store replaced.

One bookkeeping fact will save any analyst an hour of confusion: the industry changed code numbers. Under the NAICS 2017 scheme that County Business Patterns still used for 2023, gasoline stations are 447110 and 447190; under NAICS 2022, which the QCEW adopted, the same industries are 457110 and 457120. A QCEW query for 447110 in 2024 returns nothing, not because the stations vanished but because the classification moved, and any time series stitched across the boundary must stitch the codes as well.

The parking evidence: the federal survey that measures the shortage

The travel center's scarcest product is not fuel. The Federal Highway Administration's Jason's Law survey, the recurring national assessment of truck parking that Congress mandated in 2012, counted about 313,000 truck parking spaces nationally in its 2019 cycle: roughly 40,000 at public rest areas, up 6% from the 2014 survey, and 273,000 at private truck stops, up 11% (FHWA, Jason's Law Truck Parking Survey Results and Comparative Analysis, 2019 survey presented 1 December 2020). The demand side of the same survey is the remarkable half: 11,696 drivers responded, 43% more than in 2014, and 98% reported difficulty finding safe parking, with 75% experiencing it weekly. On the supply response, 79% of truck stop operators reported no plans to add parking. The shortage concentrates where the freight does: the survey's worst-shortage states were New York, New Jersey, Pennsylvania, Illinois and Georgia, and its worst corridors the length of I-95, I-5, the Chicago region and the Pacific routes. FHWA reissued the survey in 2024 with a further industry version in 2025; results had not been published as of August 2026, so the 2019 cycle remains the latest federal measurement, and this article treats it as such rather than freshening it with unsourced numbers.

For a demand model the survey does something no traffic count can: it measures the binding constraint. A combination truck's driver purchases parking with the only currency that matters, remaining legal hours under the hours-of-service rules, and a corridor where 98% of drivers report difficulty parking is a corridor where the marginal parking space captures demand almost independently of the fuel price. That is why the honest travel center model starts from the parking field and the truck-classified count rather than from retail gravity. The demand unit is a truck that must stop; the questions are how many pass, when they hit their hours, and what the corridor offers within the legal window, and two of those three are published.

Truck travel and the freight plateau

The volume base is measured annually with more precision than most analysts use. The Federal Highway Administration's Highway Statistics Table VM-1 for 2023, updated in March 2025, puts combination-truck travel at 195.8 billion vehicle miles, up 0.4% from 195.0 billion in 2022, and single-unit truck travel at 134.1 billion, down from 136.2 billion; together trucks drove 329.9 billion of the nation's 3,246.8 billion vehicle miles, 10.2% of all travel, with combination trucks at 6.0% (FHWA, Highway Statistics 2023, Table VM-1, 2025). The Bureau of Transportation Statistics' freight Transportation Services Index says the same thing monthly: annual averages of 137.3 in 2019, a pandemic dip to 131.2, then 134.3, 137.4, 137.6, 137.2, 137.5 and a 137.0 average through the first half of 2026 (BTS, freight TSI, 2026). Freight movement has been a plateau for four years, oscillating within half a point of its 2019 level.

Set the plateau against the parking evidence and the asset class's real dynamic appears, and it is this article's one finding. Travel center demand is not growing because freight is growing; freight is flat on every federal measure. The demand pressure comes from the constraint side: a fixed hours-of-service clock, a parking inventory that grew 10% in the five years to 2019 while the driver population reporting weekly difficulty stayed at three quarters, and operators, four in five, who told the federal survey they had no plans to add spaces. The scarce input at a travel center is measured in minutes of legal driving time, not gallons or tons, and it is the input the federal record measures best and the market has been slowest to build for. A corridor site's demand case built on freight growth borrows a trend that does not exist; built on the parking arithmetic of a specific corridor, it stands on the strongest evidence the public record offers.

The diesel shock of 2026, and what fuel revenue does not measure

The price series adds this year's caution. On-highway diesel averaged $3.06 a gallon in 2019, $2.55 in 2020, $3.29 in 2021, $4.99 in the 2022 spike, $4.21 in 2023, $3.76 in 2024 and $3.66 in 2025; through February 2026 the monthly averages sat near $3.60, and then the series broke upward: $4.92 in March, $5.50 in April, $5.60 in May, $5.02 in June, with the weekly average at $5.454 on 17 August 2026, up $1.741 from a year earlier (U.S. Energy Information Administration, Monthly Energy Review Table 9.4 and Gasoline and Diesel Fuel Update, 2026). This article reports the move without asserting its cause; what it insists on is the analytical consequence. Fuel is priced volumetrically, so a station's fuel revenue in mid-2026 is half again its year-ago level with no change in trucks, gallons or stops. A demand memo that quotes fuel revenue growth in a price shock is measuring the shock. The demand quantities are the truck-classified counts and the parking utilization; the price belongs in the memo as context and margin risk, and the oil infrastructure map article covers the public layer behind the price, the basins, pipelines and pricing points that decide what a corridor's diesel costs before the retail margin exists.

The geography: where stations are thick and thin

Dividing the 2023 station counts, both codes together, by the July 2023 population gives 32.1 gasoline station establishments per 100,000 residents nationally, and a state spread that maps rurality and route structure rather than wealth: Mississippi at 64.4, South Dakota at 63.2, Alabama at 60.4, Vermont at 59.4, North Dakota at 56.0 and Wyoming at 55.6 at the top; the District of Columbia at 11.9, Hawaii at 16.6, California at 20.7, Arizona at 21.4 and New York at 22.3 at the bottom (U.S. Census Bureau, County Business Patterns 2023 state file, 2025; Vintage 2025 Population Estimates, 2026; MMCG computation). The pattern is the inverse of station size: low-density states run large-format, high-volume sites, and the dense rural South and Plains run many small ones. For a travel center the state figure is context, not catchment; the asset's competition is counted along the corridor, in stations per interstate mile with truck parking fields, a count the analyst assembles from the assessor roll and aerial imagery the way every inventory in this series is assembled. What the state figure contributes is the base rate that keeps a corridor count honest, and the reminder that the format competes two ways at once, against other travel centers for trucks and against every convenience store for the four-wheel traffic that pays the inside margin.

The four-wheel side: the other half of the forecourt

Truck demand fills the diesel lanes and the parking field; the inside economics, the convenience store, the food service, the showers-and-services margin, lean equally on four-wheel traffic, and that half of the model is the ordinary retail catchment arithmetic this library covers elsewhere. The resident base within the interchange's convenience radius is read from the ACS the way the trade-area demographics article describes, households, vehicles available and commuting patterns by block group; the pass-by base is the light-duty share of the same classified count that supplied the truck volumes, which on the 2023 national figures ran 141.5 billion short-wheelbase and 53.2 billion long-wheelbase light-duty miles on rural interstates against the trucks' 71.4 billion. A travel center at a metropolitan-edge interchange is two businesses at one curb cut: an interstate service plaza for through traffic and the local fuel-and-convenience stop for three thousand nearby households, and the demand memo should carry both columns separately, because they move on different drivers and fail for different reasons. The adjacent single-purpose formats are analyzed in their own articles in this series, the car wash piece most directly, since the express tunnel increasingly shares the same outparcel and the same four-wheel count.

Corridors and seasons

Truck volumes are the steadiest series on the road, but the travel center's total demand still breathes with the calendar, because the four-wheel half does. National travel ran from 238.7 billion vehicle miles in February 2025 to 297.1 billion in July, a 24.5% summer swell, and recreational corridors run steeper than the national curve (FHWA, Traffic Volume Trends, December 2025, Table 2, 2026). For a corridor whose summer is families and RVs, the seasonal arithmetic of the campground article applies to the forecourt as well: the state's seasonal adjustment factors for the route translate the annual average into month-by-month volumes, and the inside-sales pro forma should be seasonal even where the diesel gallons are not. The reverse asymmetry is the useful one in credit: a site whose truck base covers fixed costs and debt service on its own, with the seasonal four-wheel margin as upside, carries a different risk than a site that needs July to survive February, and the classified count plus the seasonal factors are sufficient to say which one a given interchange is before anyone visits it.

Truck-classified AADT: the count that matters and the ways it goes wrong

The traffic input for a travel center is not AADT; it is the truck share of AADT, and the difference is the difference between a demand model and a wish. State traffic programs classify vehicles at their continuous and short-term count stations using the federal thirteen-class scheme, and publish, for monitored segments, the share of volume in single-unit and combination classes; the Highway Performance Monitoring System carries combination-truck and single-unit-truck AADT as data items on the Federal-Aid system, with public release geospatial files at the 2018 data year and later years through the state portals (FHWA, HPMS Field Manual and public release; state DOT programs). The state DOT traffic counts article covers where the counts come from and how estimated segments differ from counted ones; the national traffic atlas maps how much of each state's network carries a real classified count. The rural interstate figures set the scale: 141.5 billion light-duty short-wheelbase miles and 59.4 billion combination-truck miles ran on rural interstates in 2023, making the combination share of rural interstate travel roughly a quarter when the long-wheelbase light and single-unit classes are added, against 6.0% of travel systemwide (FHWA, Highway Statistics 2023, Table VM-1, 2025).

Four disciplines keep the classified count honest in a memo. State the station: a counted classification station on the site's own segment, or a modelled estimate from a station miles away, and which year. State the direction: fuel stops skew to the travel direction with easier access, and a divided rural interstate site captures its own side's trucks plus whatever share will cross for price or parking. State the time pattern: truck volumes are flatter by season than car volumes but structured by day and hour, and the hours-of-service clock makes evening truck arrivals a parking decision rather than a fuel one. And state the route logic: combination trucks on an interstate are through traffic whose stop choice is governed by spacing, the distance since the last major facility and to the next one, which is why between-stop spacing appears in every serious corridor study and why a new site's real competition may be ninety miles away, not at the next exit. All four are readable from public sources: the classification data, the directional splits, the hourly distributions at continuous stations, and the spacing from any map against the inventory the analyst has already built.

The tank registry: the inventory shortcut

Every fuel site in the country sits above registered tanks, and the registries are public. Underground storage tank programs, run by the states under the federal framework of 40 CFR part 280, maintain facility-level records of active and closed tanks with addresses, owners, tank counts and capacities, because the environmental regime requires them; the practical consequence for a demand analyst is a ready-made census of fuel infrastructure that no commercial database matches for completeness at the bottom of the market. The corridor inventory that would take a week of aerial digitizing assembles in an afternoon from the state registry filtered to the route's ZIP codes, joined to the assessor roll for parcel and improvement detail, and checked against imagery for the truck-parking field the registry does not record. The registry's closed-tank records carry their own signal: a corridor scattered with closed fuel sites marks either format consolidation, small stations absorbed by larger ones, or a demand base that left, and the assessor's dates say which. Like all environmental records, tank data serves diligence first, and nothing here substitutes for the Phase I process; the point for this article is narrower and cheerful: the most regulated feature of the asset class produced, as a byproduct, the best free supply census it has.

What the lender reads

Travel center credit concentrates three risks the public record speaks to directly. Collateral: the format is fuel infrastructure with retail attached, and underground storage tanks bring environmental diligence that sits outside this article's scope but inside every closing checklist; on classification, the special purpose examples reported from SBA guidance on SOP 50 10 8 include service formats with fixed fuel and vehicle infrastructure, and the SOP itself governs any specific property's treatment (504 Capital Corporation, 2025; NAGGL, 2025). The special purpose property article reads what the public loan record shows for single-use collateral generally. Volume: the demand quantities are the classified counts and the parking field, and the freight plateau means underwriting should assume the corridor's trucks, not the nation's growth, with the site's capture argued from spacing and access rather than from a market trend. Price: the 2026 diesel shock is margin context, and the honest pro forma separates gallons from price so a committee can see which one the revenue line assumes. The public SBA 7(a) and 504 datasets, which MMCG Analytics' SBA layer is built on, carry two decades of lending to the station and travel center codes, subject to the rule that no performance rate is shown for any cohort of fewer than ten loans. MMCG Analytics supplies the data and the analysis; the credit decision rests with the lender.

Method: the five numbers a travel center memo should carry

First, the classified count: total AADT and combination-truck AADT on the site's segment, with the station type, count year and directional split stated, from the state DOT and HPMS. Second, the spacing: distance along the route to the previous and next facilities with truck parking and diesel lanes, in both directions, with their parking counts from imagery and the assessor roll. Third, the parking arithmetic: the corridor's spaces within the legal-hours window against the overnight truck volumes the classified counts and hourly distributions imply, read against the Jason's Law state evidence. Fourth, the four-wheel base: the resident and traveling population that pays the inside margin, from the ACS for the resident side and the seasonal travel curve of the corridor for the rest. Fifth, the price context: the diesel series, annual and current, with the pro forma's gallons stated separately from its price so the demand claim survives the next move in either direction.

Each is a measurement with a named source; the model on top of them, capture rates by need (fuel, parking, food) and the revenue per stop, is the operator's forecast and should be presented as such. The boundary is the same one every article in the asset-class pillar draws and the provenance standard enforces: measured quantities carry sources and dates; assumptions carry names and stress tests; and nothing wears the other's clothes.

A worked sequence for one interchange

Run for a proposed travel center at a rural interstate interchange, the sequence goes as follows. The analyst pulls the state DOT's classified count for the segment: say 31,000 AADT with 27% combination trucks, a counted station two miles away, last full count in 2024, directional split near even. The spacing survey, run in both directions from the interchange, finds the nearest full-service facilities 38 miles east and 54 miles west, with 210 and 165 truck spaces by aerial count; two fuel-only stations sit closer but park fewer than 20 trucks each. The corridor's overnight demand is estimated from the classified volume and the hourly distribution at the state's continuous station on the same route, and set against the parking supply inside a 45-minute legal-hours window; the arithmetic shows evening deficits on the eastbound side consistent with the state's Jason's Law shortage listing. The four-wheel base is read from the ACS for the three counties the interchange serves and from the corridor's seasonal curve. The price context states 2026 diesel at its shocked level and the pro forma's gallons at the classified volumes with a stated capture range. The memo that results says: this corridor carries a counted 8,400 combination trucks a day, parks about 400 of them overnight within the window, shows a measured evening deficit, and offers spacing that makes the site the first full facility for 38 miles in the loaded direction. Whether the operator's capture assumption is conservative is now a discussion about a stated number sitting on five measured ones.

The reverse case reads as quickly. An interchange 9 miles from a 600-space flagship, on a segment whose classified count is a modelled estimate from 2019, in a state whose survey shows no shortage on that route, is a site whose demand case rests on taking share at the pump, and the memo should say so in those words before anyone prices the canopy steel.

The public side of the pipeline

Supply in this asset class has a public wing the analyst must count. Rest areas hold roughly 40,000 of the nation's truck spaces, and federal infrastructure programs have been funding state truck-parking expansions whose project lists are public documents; a corridor where the state DOT has programmed several hundred new public spaces is a corridor whose paid-parking economics and overnight deficits are about to move. The same programs fund the alternative-fuel corridor designations that will decide, exit by exit, where charging and hydrogen infrastructure lands alongside diesel. Neither this article nor the federal record forecasts that transition's pace; what the record supplies is the designation maps and the funded project lists, and a travel center memo written in 2026 that ignores them is writing the asset's second decade on the first decade's assumptions. The pipeline column of the inventory therefore carries three kinds of rows, private projects from the permit record, public parking projects from the state's program documents, and corridor designations from the federal maps, each dated, each sourced. The reading discipline is the same as everywhere in this series: a programmed public project is dated by its funding document, not treated as built; a designation is a map layer, not a certainty; and each row states what would have to happen, and by when, for the corridor's parking or fuel balance to move. A pipeline column written that way ages into a scorecard, and the analyst who kept it can show a committee not only what the corridor looks like but which of last year's expectations arrived.

The 2026 read

Read together in August 2026, the series describe an asset class whose demand story is constraint, not growth. Freight is flat: the TSI's first half averaged 137.0 against 137.3 in 2019, and combination-truck miles grew 0.4% in the latest measured year. The constraint is binding: the last published federal survey has 98% of drivers reporting parking difficulty and 79% of operators planning no additions, with the update still in the field. The station base is stable to growing: 101,351 convenience-format reporting units in 2024, up from 99,522 in 2022. And the price environment is shocked: diesel at $5.454 in mid-August, up $1.741 on the same week of 2025, which flatters every fuel revenue line it touches. For the analyst the posture follows directly: measure the corridor's trucks from the classified counts, measure its parking from the field and the survey, treat freight growth as zero until a federal series says otherwise, and keep gallons and price in separate columns. Every input named in this article publishes on a knowable schedule, the counts annually, the survey by cycle, the TSI monthly, the diesel price weekly, and a corridor position reviewed on that calendar is watched by the same instruments that justified it.

The cadence completes the method. The classified counts and VM-1 update annually, the TSI monthly with a two-month lag, the diesel price weekly on Monday afternoons, the QCEW quarterly, County Business Patterns annually, the parking survey by federal cycle, and the tank registries continuously as states process filings. A travel center position is therefore reviewable on a real calendar: price weekly, freight monthly, volume and stations annually, the constraint at each survey cycle. Few asset classes allow a credit file to be refreshed entirely from primary sources this way, and none of the refresh costs more than the analyst's time; the folklore alternative, a capture rate remembered from a conference, updates never and costs whatever the deal loses. Between those options the choice is not close, and the fact that the choice exists is this library's recurring point: the public record, read with discipline, is not the cheap substitute for market intelligence in this asset class; it is the market intelligence.

Frequently asked questions

What public data measures truck stop demand?

Four federal layers: truck-classified traffic counts from state DOTs and HPMS (combination-truck AADT); the FHWA's Jason's Law truck parking survey (313,000 spaces, 98% of drivers reporting difficulty in the 2019 cycle); freight volume series (Highway Statistics VM-1 truck miles and the BTS freight Transportation Services Index); and EIA diesel prices. The station industry itself is censused in County Business Patterns and the QCEW.

How is fuel demand modeled from AADT?

Not from AADT alone: the model starts from the truck-classified share of the count, states the station type, count year and directional split, and argues capture from spacing, the distance to the previous and next full facilities in each direction. Gallons are then stated separately from price, because a price shock like 2026's moves fuel revenue with no change in demand.

How severe is the truck parking shortage?

In the latest published federal survey (2019 cycle), 98% of 11,696 responding drivers reported difficulty finding safe parking, 75% at least weekly, against about 313,000 spaces nationally, and 79% of truck stop operators planned no additions. FHWA reissued the survey in 2024 and 2025; results were unpublished as of August 2026.

Is truck freight growing?

Not on the federal measures. The BTS freight Transportation Services Index has averaged within half a point of its 2019 level for four years (137.0 in the first half of 2026 against 137.3 in 2019), and combination-truck travel grew 0.4% in 2023 to 195.8 billion miles. Travel center demand pressure comes from the parking constraint, not volume growth.

What happened to diesel prices in 2026?

After averaging $3.66 a gallon in 2025 and holding near $3.60 through February 2026, on-highway diesel jumped to $4.92 in March, $5.50 in April and $5.60 in May, easing to $5.02 in June, with the weekly average at $5.454 on 17 August 2026, up $1.741 year over year, per EIA. Analyses should separate gallons from price when reading 2026 revenue.

How many gas stations are there in the United States?

County Business Patterns counted 96,002 gasoline stations with convenience stores and 13,172 other gasoline stations in 2023 (109,174 employer establishments in total). On the QCEW basis the convenience format grew from 99,522 reporting units in 2022 to 101,351 in 2024; the fuel-only format declined. Under NAICS 2022 the codes are 457110 and 457120.

Sources

1. U.S. Census Bureau, County Business Patterns 2023, national and state files, NAICS 447110 and 447190, released 26 June 2025. https://www2.census.gov/programs-surveys/cbp/datasets/2023/
2. U.S. Bureau of Labor Statistics, Quarterly Census of Employment and Wages, annual averages 2022 and 2024, NAICS 457110 and 457120, private, United States, 2023 and 2025. https://data.bls.gov/cew/data/api/2024/a/industry/457110.csv
3. Federal Highway Administration, Jason's Law Truck Parking Survey Results and Comparative Analysis, 2019 survey results presented to the National Coalition on Truck Parking, 1 December 2020; 2014 survey per the 2015 report. https://ops.fhwa.dot.gov/freight/infrastructure/truck_parking/jasons_law/truckparkingsurvey/jasons_law.pdf
4. Federal Highway Administration, Jason's Law survey reissue, 2024 and 2025 cycles (results not yet published as of August 2026). https://ops.fhwa.dot.gov/freight/infrastructure/truck_parking/index.htm
5. Federal Highway Administration, Highway Statistics 2023, Table VM-1, Annual Vehicle Distance Traveled in Miles and Related Data, updated March 2025. https://www.fhwa.dot.gov/policyinformation/statistics/2023/vm1.cfm
6. U.S. Bureau of Transportation Statistics, Transportation Services Index, freight series, monthly through June 2026 (data.bts.gov series bw6n-ddqk); annual averages computed by MMCG. https://data.bts.gov/resource/bw6n-ddqk.json
7. U.S. Energy Information Administration, Monthly Energy Review, Table 9.4, On-Highway Diesel Fuel Price (series DFONUUS), monthly and annual through June 2026. https://www.eia.gov/totalenergy/data/browser/csv.php?tbl=T09.04
8. U.S. Energy Information Administration, Gasoline and Diesel Fuel Update, weekly on-highway diesel prices for 3, 10 and 17 August 2026. https://www.eia.gov/petroleum/gasdiesel/
9. Federal Highway Administration, Highway Performance Monitoring System Field Manual and HPMS Public Release (combination-truck and single-unit-truck AADT data items), page updated 13 September 2022. https://www.fhwa.dot.gov/policyinformation/hpms/shapefiles.cfm
10. U.S. Census Bureau, Vintage 2025 Population Estimates, state totals, July 2023 populations, 2026. https://www2.census.gov/programs-surveys/popest/datasets/2020-2025/state/totals/NST-EST2025-ALLDATA.csv
11. 504 Capital Corporation, SBA 504 Loans for Special Purpose Properties and Real Estate (guidance on SOP 50 10 8 classifications), 2025. https://504capital.com/blog/financing-special-purpose-properties-sba-504-loans/
12. National Association of Government Guaranteed Lenders, SBA Notice Revising SOP 50 10 8 (Procedural Notice 5000-872764), 2025. https://www.naggl.org/sba-notice-revising-sop-50-10-8/
13. U.S. Environmental Protection Agency, Underground Storage Tank program regulations, 40 CFR part 280 (state-run facility registries of active and closed tanks), current edition, 2025. https://www.epa.gov/ust
14. MMCG Research, SBA 7(a) Performance Series: MMCG analysis of the public SBA 7(a) loan register, 2026. https://mmcganalytics.com/sba-default-rates/