US Gas Station Industry Outlook 2026: Peak Demand, Record Margins
- 11 minutes ago
- 28 min read
By Michal Mohelsky, JD, principal, MMCG Invest, LLC | July 25, 2026
Gasoline peaked in 2018. So did the American gas station, if you count them. And yet 2025 was, by the industry's own accounting, one of the most profitable years fuel retail has ever had. Both things are true, and the distance between them is this report.
Here is the shape of the paradox. U.S. gasoline demand topped out at about 9.33 million barrels a day in 2018 and has never returned; 2025 ran roughly 4 percent below the pre-pandemic level (5). The store count peaked the same year, at 154,958, and has drifted down to 151,975 (3). Meanwhile the industry sold $817.5 billion of goods and fuel in 2025, posted its twenty-third consecutive record year of in-store sales, and earned fuel margins above 40 cents a gallon, roughly double the pre-2020 norm (2). Convenience and gas real estate now trades at the tightest cap rates in all of retail (21). The sector attracted the largest attempted retail takeover in history, watched it collapse, and then produced three IPO filings inside twelve months (17, 14, 19).
The resolution of the paradox is simple to state and has enormous consequences for anyone lending against or investing in this asset class: the gas station stopped being a fuel business. Fuel still generates 65 percent of the sales dollars but only 38.8 percent of the gross profit dollars (2). The store ate the pump. Foodservice, packaged beverages and the real estate underneath now carry the economics, and the gallon has been demoted to a customer acquisition tool that happens to gross forty cents a unit.
That inversion changes what the asset is worth, who should own it, and how a 25-year loan written in 2026, maturing in 2051, should be underwritten. This report walks through the whole machine: the P&L, the supply churn, the margin question, the demand curve, the regulatory map, the deal market, the SBA loan tape, the tank in the ground, and the food program that decides who survives. Where a claim rests on our own analysis of federal loan-level or environmental data, we say so. Where a widely repeated industry number could not be traced to a primary source, we say that too.
Key Statistics: The U.S. Gas Station and Convenience Industry in 2026
Total industry sales reached $817.5 billion in 2025, down 2.4 percent on lower pump prices even as gallons and inside sales grew (2).
In-store sales hit a record $341.2 billion, the twenty-third straight annual increase (2).
Fuel was 65.0 percent of sales dollars but only 38.8 percent of gross profit dollars (2).
Store count: 151,975 at the end of 2025, down from the 2018 peak of 154,958 (3).
Fuel-selling stores rose to 122,620, an eight-year high; 80.7 percent of c-stores sell fuel (3).
About 63 percent of stores belong to operators with ten or fewer locations; roughly 60 percent are single-store operators (3).
Retail fuel margins averaged above 40 cents per gallon in 2025 versus roughly 22 cents pre-2020; adjusted for inflation, about half of that doubling disappears (4, 1).
Card fees hit a record $21.3 billion industrywide, including $4.6 billion assessed on fuel taxes the retailer merely collects (4).
Net fuel margin after expenses runs about 10 to 15 cents per gallon, per NACS's own arithmetic (4).
U.S. gasoline demand peaked in 2018 at about 9.33 million barrels per day; 2025 averaged about 8.9 million (5, 6).
EV share of new vehicle sales collapsed from a record 10.5 percent in Q3 2025 to 5.8 percent after the federal credit expired September 30, 2025 (29, 30).
A representative new suburban store costs about $6.1 million all-in to build; a Buc-ee's runs $47 to $50 million (1).
A new build needs roughly 3.5 to 4 million gallons a year, three to four times the legacy-store average, to pencil (1, 15).
Couche-Tard withdrew its roughly $47 billion bid for Seven & i on July 16, 2025, the largest failed retail deal ever (17).
Single sites trade at roughly 2.5 to 4.5 times earnings; scaled platforms fetch high single to low double digits; Casey's trades near 25 times EBITDA (1, 16).
Wawa net lease deals priced at 4.90 to 5.20 percent cap rates in Q1 2026, the tightest in retail (21).
Gas stations are the fourth-largest SBA 7(a) category ever, with $17.7 billion lent since 1995 and another $6.5 billion across 4,166 loans just since fiscal 2020; the lifetime dollar charge-off rate is 3.5 percent while the resolved-loan default rate runs near 15 percent (31, 32, 1).
533,277 active underground storage tanks remain at about 190,000 facilities; roughly 275,000 are already 30 or more years old (8, 9).
Average tank-release cleanup costs $142,219, and the number of million-dollar cleanups doubled between 2014 and 2018 (10).
Foodservice is 28.5 percent of in-store sales but 38.9 percent of in-store gross profit; for the top decile of stores it is 46.6 percent of inside profit, for the bottom decile just 8.3 percent (2, 19).
1. The $817 Billion Machine: How the Money Actually Flows
Start with the number everyone quotes and almost everyone misreads. Total industry sales fell from $906 billion in 2022 to $817.5 billion in 2025 (2). That looks like decline. It is mostly arithmetic: about two thirds of the top line is fuel priced at the pump, so when average gasoline fell from $3.30 to $3.11 a gallon, the industry "shrank" while selling more gallons and more merchandise than ever. Revenue is the wrong lens for this business. Gross profit is the right one.
On the profit lens, the inversion is complete. Fuel produced 65.0 percent of 2025 sales dollars and 38.8 percent of gross profit dollars (2). The inside of the store, 35 percent of sales, produced 61 percent of the profit. Within the inside, foodservice alone was 28.5 percent of sales and 38.9 percent of gross profit (2). Prepared food, at margins around 55 percent, would rank as the single largest inside category on its own. Cigarettes, the historical anchor, fell to third in profit contribution, and in 2025 other tobacco products, led by nicotine pouches, passed cigarettes in gross profit per store for the first time (4, 36).
Two stacked bars, "Where the dollars come from" vs "Where the profit comes from," 2025: fuel 65.0 / 35.0 inside on sales; fuel 38.8 / 61.2 inside on gross profit. Source 2.
Two structural facts about this machine matter for underwriting. First, the averages hide brutal dispersion. NACS's own historical analysis found only the top quartile of firms could turn a profit on inside sales alone; for everyone else, fuel margin subsidizes an unprofitable box (2, 1). The gap between top-decile and bottom-decile foodservice performance is 7.6 to 1 in sales and wider in profit. Second, the cost side is squeezing. Store wages average $15.04 an hour across 19.9 employees per store, direct operating expenses rose 4.2 percent in 2025 (the slowest since COVID, which tells you about the prior years), and card fees, the second-largest expense after labor, set another record at $21.3 billion (2, 4).
An average store now rings 45,160 transactions a month, and inside transaction counts fell 1.6 percent in 2025 (2). The industry is not defending visit counts. It is defending profit per visit. That is the entire strategic story of the next decade, and it is why the rest of this report keeps returning to one question: what, exactly, gets people to stop when the tank no longer requires it?
2. Fewer Stores, More Pumps: Supply Is Reconstituting
The store count headline, down about 3,000 stores from the 2018 peak, suggests gentle decline. Underneath it, supply is not shrinking. It is being rebuilt into a different shape: small, old, under-scaled boxes are dying while very large, food-forward sites are going up at a pace the trade press undersells.
The churn is visible in one pairing: total stores fell 280 in 2025 while fuel-selling sites rose 768 to 122,620, the highest in eight years (3). 7-Eleven alone is running the sector's largest rationalization, 444 North American closures announced in October 2024 and another 645 sites leaving the store network in fiscal 2026 through closures, wholesale conversions and franchise terminations, with roughly 2,600 company stores headed to franchisees by 2030 (35). Its stated reasons read like the industry's diagnosis of itself: traffic down, inflation, and cigarette volumes off 26 percent since 2019, an eighty-year low (35).
On the build side, the pipelines are named and dated. Buc-ee's operates 56 stores in 13 states, is entering six to eight more, and its 74,000-square-foot, 120-position prototype costs $47 to $50 million (1). Casey's is targeting at least 400 stores over fiscal 2027 to 2029, split between builds and acquisitions, noting an acquired store costs about $1 million less than a new one at the same return hurdle (16). Murphy USA is running 45 to 55 new-to-industry stores a year plus up to 30 raze-and-rebuilds (15). Love's budgets $700 million for 20 travel centers in 2026. Wawa, Sheetz and QuikTrip are all crossing state lines they had never touched.
What does entry cost? Our own itemized example, an eight-position, 3,024-square-foot store on 5.4 acres, totals $6,067,800 all-in: land 12 percent, hard construction 52 percent, equipment 19 percent, soft costs 9 percent, financing 8 percent (1). Vendor and permit data bracket the range: roughly $2.5 to $6 million for a standard suburban store, $10 to $15 million for a travel center. Tariffs made it worse; steel mill products rose 20.7 percent and aluminum 33 percent in the year to early 2026, and a forecourt is unusually metals-intensive (1).
Here is the arithmetic that explains why chains build big. A $6 million store needs roughly $600,000 to $720,000 of NOI for a 10 to 12 percent yield on cost, which at realistic margins means 3.5 to 4 million gallons a year plus several million dollars of inside sales (1). The industry-average legacy site pumps about a million gallons. Murphy's recent new-store classes run over 310,000 gallons a month (15). New supply is not competing with the average store. It is a different species, and when one lands within two miles of an incumbent, the incumbent's fuel volume is at genuine risk even though nobody has published a clean study quantifying it (40). The entitlement barrier cuts the other way: conditional use permits, distance separations, hostile hearings and 12-to-24-month timelines make replicating an existing entitled, tanks-in-the-ground corner genuinely hard, which is quiet support under the value of the incumbent asset.
Dual line, 2016-2025: total store count (154,195 to 151,975, peak 154,958 in 2018) vs fuel-selling sites (rising to 122,620). Source 3.
3. The Forty-Cent Question: Is the Margin Real?
Everything in a fuel-site pro forma hangs on one assumption: the cents-per-gallon margin. Before 2020 the national gross margin ran about 22 to 24 cents. In 2025 it averaged above 40 (4). The trade now repeats "40 is the new 20" as settled fact, and Yesway's IPO prospectus argues in writing that margins have "settled at a durable new baseline" (14). A lender writing 25-year paper needs to know whether that is a repricing or a cycle.
Our answer, after working both sides of the argument: it is partly real, and the part that is real is smaller than it looks. Deflate the series and roughly half of the nominal doubling is just inflation; the real gain since 2019 is closer to 40 percent (1). The mechanisms that hold up under scrutiny are cost-driven, not power-driven. Declining gallons per site and labor costs that rose from about $80,000 to $104,000 per store per month over five years force weaker operators to price up, and scaled players shelter under that umbrella; Murphy USA says as much in its annual report, and its own retail margin has sat at a stable 28.1 cents for two straight years (15, 4). The market-power stories are weaker. The industry is still 63 percent mom-and-pop, drivers still detour five minutes to save five cents, fueling sites are increasing rather than exiting, and pricing software is now the subject of a California algorithmic-pricing statute and live litigation (3, 25).
Then there is the drag nobody prices: cards. Fees are charged on the full transaction value including the taxes, so the industry paid $4.6 billion in 2025 in fees on money it merely collects for governments (4). At roughly 8.4 cents a gallon, interchange consumes a fifth to a quarter of the gross margin, and it rises mechanically with pump prices. Walk the whole waterfall the way NACS does and the 40-cent headline lands at 10 to 15 cents of net margin, three to five percent on a three-dollar gallon (4).
Cents-per-gallon waterfall, 2025: gross 40.0, card fees -8.4, distribution -6.0, store-level fuel opex -12.6, net ~13.0. Sources 4, 1; sub-allocations MMCG estimates.
The underwriting conclusion we give credit committees: base-case a single site at 28 to 32 cents gross, not 40; deduct card fees as a percentage of the projected pump price, not a fixed number; stress the worst year back to the low 20s; and treat any trailing margin above the low 30s as cyclical upside you do not capitalize (1). The counterintuitive corollary is that falling gasoline prices, which the EIA projects into 2027, are good for this industry: margins fatten on the way down and the card-fee drag shrinks with the price (6).
4. Peak Gallon: Demand After 2018
The gallon curve is the slowest-moving variable in the model and the most argued about. The facts first. U.S. finished gasoline product supplied peaked in 2018 at about 9.33 million barrels a day, collapsed in 2020, recovered to about 8.9 million by 2024-2025, and per the EIA will keep sliding about one percent a year even as miles driven exceed pre-pandemic levels, because the fleet keeps getting more efficient (5, 6, 27, 28). Note what is not driving that decline yet: electric cars. EVs are still only about two percent of the vehicles actually on the road, and the fleet's average age hit a record 12.8 years, so new-vehicle efficiency seeps into gallons slowly (28).
Then 2025 happened to the EV curve. The federal purchase credit terminated September 30, 2025; sales spiked to a record 10.5 percent share in the third quarter as buyers beat the deadline, then fell to 5.8 percent in the fourth and stayed there into 2026 (29, 30). The same July 2025 law zeroed out CAFE penalties, and Congress revoked California's Advanced Clean Cars II waivers in June 2025, suspending the state mandates that would have forced the steepest gallon decline (30). The forecasting establishment moved with it: the EIA's 2026 long-term outlook now spans an 11 to 23 percent petroleum decline by 2050 depending on whether standards survive, and the IEA, which two years ago called peak oil by 2030, now shows demand rising to 2050 in its current-policies case (7). When the referees keep moving the goalposts in both directions, a 25-year point forecast is a scenario, not a base case.
What a lender can actually use is geography. EV share of new sales spans roughly 23 percent in California to under one percent in Mississippi, Wyoming and North Dakota, a twenty-fold spread (1). Our overlay: underwrite gallons at minus 0.3 to 0.8 percent a year at rural, diesel-heavy, low-EV sites; minus 1 to 1.5 percent on typical suburban corridors; minus 2 to 4 percent, stressed to 5, in dense high-EV coastal metros (1). Diesel and the travel corridor are the insulated end; freight electrifies slowly and truck miles keep growing (27).
And the charger? At mid-2026 economics it is a real estate bet, not a margin replacement. National DC fast-charging utilization ran about 16.4 percent against a roughly 20 percent break-even, installed cost reaches into the low hundreds of thousands per port, and demand charges bury low-volume sites (1). Only 1.4 percent of c-stores offer charging. Credit it in a pro forma only where utility make-ready or grants cover the capex and the site can prove the dwell time turns into basket.
Line 2005-2025: gasoline product supplied (peak 9.33M b/d 2018, 8.9M 2025), with inset bar of quarterly EV sales share 2024-Q1 2026 showing the 10.5 to 5.8 whiplash. Sources 5, 29.
5. Radically Local: Taxes, California, and the Ban That Helps You
There is no such thing as the U.S. fuel market. There are fifty of them. State taxes and fees on gasoline ran from 9.0 cents in Alaska to 70.9 cents in California as of January 1, 2026, before the fixed 18.4-cent federal excise that has not moved since 1993 (22). Retail prices in July 2026 spanned roughly $3.35 in Indiana to $5.50 in California. Breakeven volume, cross-subsidy math and even the legality of your pricing (Wisconsin mandates a minimum markup; New Jersey still bans self-service) all change at the state line (22).
California deserves its own chapter because it is running the whole bear case at once. The state lost about 17 percent of its refining capacity as Phillips 66 shut Wilmington and Valero wound down Benicia by early 2026, layered a low-carbon fuel standard and a cap-and-trade program worth on the order of 23 to 27 cents a gallon onto the pump, and, most relevant to this report, enforced the country's first hard tank mandate: every single-walled underground tank had to be permanently closed by December 31, 2025, no exceptions (24, 25, 23). Roughly 50,000 single-wall tanks were affected and 99 percent were closed by the deadline, with about 600 shut in December alone (23). It is the clearest case in America of regulation directly forcing station attrition, and it is a template other states can copy.
The municipal story is noisier than it is large. Petaluma, California became the first U.S. city to ban new gas stations in February 2021, and roughly fifteen named jurisdictions have followed, overwhelmingly clustered in the Bay Area, Napa and Marin, plus outliers in Colorado and Tennessee (26). Together they hold well under one percent of the U.S. population, and Florida has preempted such bans outright. Treat it as a submarket risk, not a national trend.
Here is the finding that surprises people: for the incumbent, the ban is a gift. A jurisdiction that prohibits new stations has frozen the competitive set for the life of your loan. Combined with California's forced retirement of thousands of non-compliant competitors, a modern, double-walled, fully entitled site inside a restrictive jurisdiction carries a scarcity moat that partially offsets the demand and cost headwinds. In our underwriting work we now treat a defensible entitlement in a restrictive market as a credit positive, provided the subject site itself is already compliant.
6. Who's Buying: A $47 Billion No and an IPO Wave
If the asset were dying, capital would be leaving. It is doing the opposite, loudly. The completed benchmark remains 7-Eleven's $21 billion purchase of 3,800 Speedway stores in 2021, still the largest c-store deal ever (35). The defining event since was a failure: Alimentation Couche-Tard's pursuit of Seven & i, raised to about $47 billion, would have been the largest foreign takeover of a Japanese company in history; Couche-Tard walked away on July 16, 2025, citing obstruction, and Seven & i pivoted to asset sales, buybacks and a planned U.S. listing of 7-Eleven North America (17). The consolidation kept moving underneath: Sunoco closed its $9.1 billion Parkland acquisition on October 31, 2025; Casey's paid $1.145 billion for 198 CEFCO stores at about 11 times adjusted EBITDA; Couche-Tard took GetGo for $1.6 billion (18, 16). And the public market reopened: ARKO carved out its wholesale arm in a February 2026 IPO, Yesway priced in April 2026 at a $1.2 billion valuation, and EG Group filed its U.S. business, rebranded Cumberland Farms, in July 2026 at a reported valuation above $9 billion (20, 14, 19).
Why does money keep coming to a sector whose core commodity has peaked? Because the buyers are not underwriting gallons. They are underwriting elevated fuel margins, growing food profit, a fragmented base of 92,000 single-store operators available for roll-up, land that qualifies for 100 percent bonus depreciation since July 2025, and inflation-protected rent (2, 3, 21). The dirt tells the same story: convenience and gas is the tightest-priced category in net lease, with Wawa deals at 4.90 to 5.20 percent caps against a 6.80 percent all-sector average (21).
The number that matters most to our readers sits at the bottom of the ladder. A single independent station sells for roughly 2.5 to 4 times its earnings without real estate, 4 to 7 times with it (1). A fifty-store platform commands high single to low double digits. Casey's trades near 25 times (16). Assemble twenty stores at 4 times, roughly $20 million for $5 million of store-level EBITDA, and the platform is worth $45 to $55 million to the next buyer before you improve a single store. That multiple arbitrage, buy small, professionalize, sell scaled, is the sector's core wealth-creation machine, and it is exactly what Yesway's sponsors just monetized in public. It is also what Mountain Express tried with $205 million of debt and no integration discipline; it went from 828 fueling centers to Chapter 7 in five months of 2023 (37). The ladder rewards the climb and punishes the leverage.
Multiple ladder bars: single site business-only 2.5-4x; single site fee simple 4-7x; 51-200 store platform ~9-11x (CEFCO 11x marker); public compounders 10-25x (Casey's ~25x). Sources 1, 16.
7. The Loan Tape: What SBA Data Actually Says
Gas stations carry a folk reputation in credit circles: high-default, environmentally cursed, special-purpose trouble. We pulled the record. Gasoline stations with convenience stores are the fourth-largest SBA 7(a) category in history, $17.7 billion across 21,149 loans since 1995 (31). The performance answer depends entirely on the denominator, and the industry habitually mixes them. Measured as dollars charged off against dollars approved, the sector's lifetime rate is 3.5 percent, better than restaurants (31). Measured as defaults against resolved loans, the rate runs near 15 percent, and our own Feasibility Index, built on the same federal files, puts the historical figure near 10 percent, several times the program-wide cohort average (32, 1). Both numbers are true. Frequency is high; severity is low, because when these loans fail, the land, building and tanks recover most of the balance. The practical lesson is structural: long, real-estate-secured, 504-anchored paper has defaulted at a fraction of the rate of short goodwill-heavy loans, and operator quality dominates everything else (1).
The files themselves say more, so we ran the tape. The SBA's loan-level records current to March 31, 2026 show 4,166 gas station 7(a) approvals for $6.5 billion since fiscal 2020, the average loan climbing from $1.4 million to $1.8 million, a stimulus-driven spike in 2021, and 680 loans in fiscal 2025 alone; add the 504 program, which just wrote its two largest gas station years in the sixteen-year file in 2024 and 2025, and the sector is drawing roughly $1.3 billion of SBA credit a year (1). Performance on that fresh tape beats the folklore. Of roughly 3,400 gas station 7(a) loans disbursed since fiscal 2020, fifteen have charged off, about 0.4 percent, against 1.9 percent across all industries in the same young cohort (1). The mature 504 vintages cut the other way and complete the picture: gas station loans approved 2010 through 2016 charged off at 3.3 percent of disbursed loans against 1.8 percent program-wide, elevated frequency inside the safest structure, yet dollars lost were 2.2 percent of dollars approved over a decade and a half (1). And the sector already borrows the safe way: seven in ten of its 7(a) loans carry terms of 23 years or longer, the median is exactly 25, and in the 2020 to 2022 vintages the long real-estate paper is showing a fraction of the distress of the shorter loans (1).
The lending itself is a specialist's market. Since fiscal 2020 the busiest gas station lenders on the federal files are Celtic Bank, Readycap, Metro City, Wallis, Commonwealth Business and Open Bank, with Byline and Huntington close behind, and the sector's median initial rate ran 9.25 percent in fiscal 2025 against Prime-plus spreads that price it inside the national 7(a) median (1, 33). The folk geography, meanwhile, is out of date: Texas now leads with 20 percent of the deals since 2020, California follows at 16, and together they hold barely more than a third of the market, not the two thirds sometimes claimed, with Washington, Georgia and the Midwest taking most of the rest (1). Roughly a quarter of the loans finance a change of ownership and another third fund startups and new stores, which is precisely the acquisition-and-build economy this report keeps describing (1).
What actually governs closing a file is the SOP, and the current edition, SOP 50 10 8, effective June 1, 2025 with technical updates in March 2026, treats gas stations as one of only three named special-use facility types (11). The consequences: a Phase I environmental site assessment is mandatory regardless of loan size; contamination stops approval and disbursement unless a defined mitigant is documented, meaning state trust fund coverage, a no-further-action letter, an indemnification, or an escrow of 150 percent of remediation cost; the appraisal must be a going-concern report allocating land, improvements, equipment and intangibles; equity injection came back to a hard 10 percent on changes of ownership, with seller notes countable only on full lifetime standby and capped at half the injection (11). FY2026 fees run 2 to 3.75 percent by tranche (13). And one genuinely expansive change: as of July 4, 2026, the SBA decoupled the 7(a) and 504 caps, allowing up to $10 million of combined exposure on one project, sequence the 7(a) first (12). For a $9 to $12 million large-format build or a multi-site acquisition, the financeable envelope just doubled.
Grouped bars: the two default measures (3.5% dollar charge-off vs ~15% resolved-loan) against all-industry comparators, plus a line of sector 7(a) volume by fiscal year. Sources 31, 32, 1.
8. The Tank Clock
Every fuel site sits on a buried, regulated, liability-bearing asset with a finite life, and almost nobody in the trade press writes about it. We consider it the single most under-priced variable in the asset class.
The counts first. The country operates 533,277 federally regulated underground storage tanks at about 190,000 facilities, down from over two million in the late 1980s (8). The 1988 federal rule gave every pre-existing tank until December 22, 1998 to upgrade or die, which triggered a massive replacement wave and created a single giant vintage: tanks installed in the mid-to-late 1990s, now 28 to 32 years old. Fiberglass and protected-steel tanks carry 30-year warranties; the EPA and state regulators peg the average life of a removed tank at about 20 (9). The agency's own arithmetic says roughly 275,000 operating tanks are already past 30, nearly 400,000 will be by 2032, and more than 100,000 will come out of the ground by 2034 (9). That is not a risk forecast. That is a scheduled national capital event.
What happens when tanks come out? California, which just forced the issue, gives the cleanest data in existence: closures of single-walled systems revealed previously unknown contamination 29.7 percent of the time, versus 6.7 percent for double-walled (9, 23). Nationally, 581,676 releases have been confirmed since the program began, 53,777 remain open, and while new releases are near record lows, the EPA expects discoveries to rise as the old cohort exits (8). The money: the average completed cleanup costs $142,219, and the population of cleanups exceeding one million dollars doubled between 2014 and 2018 (10). A full three-tank replacement at an operating site runs roughly $500,000 to $1.5 million including the weeks of lost business (1).
Who pays is the lender's question. Thirty-six states run cleanup trust funds, which the SBA explicitly accepts as a contamination mitigant, but the funds are aging along with the tanks: three have closed to new releases, several are thinly funded, and California's sunsets in 2036 and can no longer be an owner's sole financial assurance (11, 23). A fund that is closed or insolvent is not a mitigant, and private tank insurance is repricing hard at the 30-year mark. The lender's own shield, the federal secured-creditor exemption, survives only if the bank never participates in managing the property and divests promptly after foreclosure (39).
Our tank-clock rule for credit files: tank age, material and wall type are underwriting fields, not appendix trivia. Past 25 years, require the replacement plan; past 30 or single-wall, fund the reserve, roughly the replacement cost plus the probability-weighted cleanup, or decline. A 25-year loan closing this year will see the 1998 cohort replaced at least once before maturity. Someone's capital does that. The pro forma should say whose.
[CHART 7 HERE. Area chart of active USTs 1990-2025 (2.1M to 533k) with age-wave overlay: 275k tanks 30+ today, ~400k by 2032; secondary line of open LUST backlog to 53,777. Sources 8, 9.]
9. Food Is the Exit: Does the Store Actually Pencil?
Everything above says the store must carry the site. So the final question is whether a food program actually works for the people who borrow from our readers: operators of one to ten stores who cannot build a commissary.
The scale numbers are seductive. Foodservice went from 11.9 percent of inside sales in 2005 to 28.5 percent in 2025 and throws off 38.9 percent of inside profit at margins around 55 percent (2). But that 55 is a gross margin before labor and before waste, and the dispersion behind the average is the whole story: the top decile of stores earns 46.6 percent of its inside profit from food, the bottom decile 8.3 percent (2, 19). And the tailwind has stalled; adjusted for inflation, c-store foodservice sales actually fell in 2025 as the burger chains went to war on price, with $5 and $4 bundle menus resetting the customer's anchor, and the drive-thru coffee chains, 7 Brew up 87 percent in units in a single year, attacked the most profitable beverage in the store (4, 1).
Our break-even arithmetic for a single site is sobering. A modest hot program with no hood, licensed pizza plus a roller grill, adds roughly $60,000 to $67,000 a year of incremental labor, waste, depreciation and utilities, which requires about $17,000 to $19,000 a month of food sales just to break even; a made-to-order kitchen adds a $40,000-plus hood and pushes the bar past $30,000 (1). Chain averages sit above these lines. Many rural single sites do not.
That is why the realistic path for the small operator is the licensed, no-fee model. Hunt Brothers Pizza and Krispy Krunchy Chicken charge no franchise fee, no royalty and no ad fund, supply the equipment, and fit in as little as 59 square feet, which is precisely why they blanket the independent channel while true franchises layer 5 to 8 percent royalties onto a margin that cannot absorb them (1). The other proven rural play is the hero item. Allsup's built an empire on one proprietary deep-fried burrito, 27 million a year across 441 mostly rural stores, and that focus, one high-repeat item instead of a broad menu bleeding waste, is the model a low-traffic site can actually execute (14).
For credit officers reviewing a borrower's food projection, the red flags are consistent: chain-average sales applied to a single site, margins quoted before labor and waste, no waste line at all (assume at least 15 percent), no hood capital in a hot-food build, delivery revenue booked at in-store margin against 25 to 30 percent platform commissions, and a franchise royalty that never made it into the model (1).
Line 2005-2025 of foodservice share of inside sales (11.9 to 28.5) and of inside gross profit (to 38.9), with decile markers 46.6 vs 8.3. Sources 2, 19.
Opportunities
The clearest money in this sector is bought, not built. Independent product routinely trades below its $4 to $6 million replacement cost while entitled corners get harder to replicate every year, so the acquisition-plus-improvement path carries a margin of safety the new build does not (1). The multiple ladder does the rest: every credible step from one store toward ten re-rates the same cash flow. Food-forward conversion is the operating lever, and the licensed no-royalty programs let a small operator take it without franchise drag. Diesel-weighted corridor sites offer the most durable demand pool on a 25-year horizon. Bonus depreciation makes fee-simple deals unusually tax-efficient right now, and the new $10 million combined SBA envelope finances projects that were un-bankable in 2024 (12, 21). And in the handful of jurisdictions that banned new stations, the incumbent owns a legislated moat.
Risks
Rank them honestly. First, margin reversion: the difference between underwriting 40 cents and 30 cents is the difference between a 1.4x and a sub-1.1x coverage on many files, and the cyclical layer above the real floor will mean-revert eventually (1, 4). Second, the tank clock and the quiet erosion of the state funds that are supposed to stand behind it. Third, card fees, which rise with pump prices and compound outside the operator's control unless Congress or the courts intervene (4). Fourth, competitive obsolescence: a large-format entrant two miles away is a bigger threat to a legacy site than any EV curve this decade. Fifth, policy whiplash in both directions on electrification; the 2025 reversals slowed the gallon decline, and a future reversal of the reversals would speed it back up, state by state (29, 30). Sixth, the QSR price ceiling on food, the one growth engine everyone is counting on. And last, the human one: sixty percent of this industry is a single owner with keys and no succession plan.
Outlook to 2031
Our base case is unglamorous and, we think, resilient. Gallons decline about 1 to 1.5 percent a year nationally, faster on the coasts, barely at all in the diesel belt (1, 6). Gross fuel margins settle in the high 20s to low 30s in real terms, still well above the old world, with cyclical spikes lenders should enjoy and not capitalize. The store count drifts toward 150,000 while fuel-selling sites hold, and the composition keeps shifting: more 5,000-square-foot food boxes, fewer 1,200-square-foot cigarette boxes. Consolidation resumes from the 2025 pause; the Cumberland Farms and 7-Eleven North America listings, if completed, will give the private market two more public marks and probably pull more sellers off the fence (19, 35). More than 100,000 tanks come out of the ground by 2034, forcing the deferred-capex reckoning this report keeps circling (9). Foodservice growth returns to low single digits and keeps deciding, store by store, who is in the top decile and who is subsidizing an unprofitable box with fuel margin.
The 2051 question we opened with has a plain answer. The gallon will still exist at maturity, smaller and geographically uneven, but it was never the collateral. The collateral is the corner, the entitlement, the tanks and their compliance record, and a store that people visit on purpose. Underwrite those four things and the asset class is, by the federal government's own loan data, a low-severity credit. Underwrite the gallons and you are lending against a customer-acquisition expense.
Frequently Asked Questions
Are gas stations a dying business? The commodity peaked; the business did not. Industry gross profit is at record levels because fuel margins doubled nominally and the inside of the store, especially food, now produces 61 percent of profit dollars (2). The asset is transitioning, not dying, and the transition has clear winners and losers.
What is a single gas station worth in 2026? Without real estate, roughly 2.5 to 4 times owner earnings plus inventory and fuel at cost. With the land, 4 to 7 times. National median fee-simple value sits near $450,000, with high-volume food-forward sites worth multiples of that (1). The same cash flow inside a 50-store platform is worth two to three times more per dollar, which is the whole roll-up thesis.
What does it cost to build a new station? About $2.5 to $6 million for a standard suburban store all-in (our itemized example: $6.07 million), $10 to $15 million for a travel center, and $47 to $50 million for a Buc-ee's-scale destination (1). A new build needs three to four times the average store's fuel volume to justify the cost.
Is the 40-cent fuel margin permanent? Partly. Adjusted for inflation, about half the increase since 2019 is real, supported by cost inflation and the fixed-cost math of declining gallons. We underwrite 28 to 32 cents and treat anything above the low 30s as cyclical (1, 4).
Will EVs strand the asset before a 25-year loan matures? Not uniformly. EVs are about two percent of the on-road fleet, adoption fell by half after the federal credit expired, and the exposure is radically state-specific, roughly 23 percent of new sales in California versus under one percent in Mississippi (29, 1). Underwrite the site's state and format, not the national headline.
What is the real SBA default rate for gas stations? Both published numbers are true. Lifetime dollar charge-offs run 3.5 percent, below restaurants, because real estate recovers well. Default frequency on resolved loans runs 10 to 15 percent, above the program norm. On the fresh federal files, gas station 7(a) loans written since 2020 have charged off at about a quarter of the program-wide rate so far. Long, real-estate-secured structures with experienced operators sit at the good end of every measure (31, 32, 1).
What kills a gas station loan? In order of frequency: environmental findings without a mitigant, single-wall or aged tanks, a ground lease with no real estate collateral, thin equity, a restrictive fuel supply agreement expiring mid-loan, and cash books that cannot be verified (11, 1). The SOP makes the first one non-negotiable: contamination stops disbursement.
MMCG Invest, LLC prepares bank-ready feasibility studies for gas station and convenience store acquisitions, construction and SBA/USDA financings. If you are underwriting one of these assets, we would be glad to help: mmcginvest.com.
July 25, 2026 Author: Michal Mohelsky, J.D., Principal, MMCG Invest, LLC
Request a Feasibility Study https://calendar.app.google/EJzWEz3GCqLY2jU86
Michal Mohelsky, J.D. | Principal | mmcginvest.com
Contact: michal@mmcginvest.com
Phone: (628) 225-1125
Disclaimer: This report is provided for informational purposes only and does not constitute investment, legal, or tax advice. Data presented herein is derived from proprietary MMCG databases and third-party sources believed to be reliable; however, MMCG Invest makes no representation as to the accuracy or completeness of such information. Figures from third-party industry databases have been independently verified and, where appropriate, adjusted to reflect MMCG's proprietary analytical methodology. Statutory and regulatory references are provided for context and must be verified with counsel before reliance. Past performance is not indicative of future results.
Sources
(1) MMCG Invest, LLC proprietary database; MMCG tabulation of the SBA FOIA loan-level files (7(a) approvals FY2020 through FY2026 and 504 approvals FY2010 through FY2026, both as of March 31, 2026, NAICS 447110, 447190, 457110 and 457120); MMCG Feasibility Index; MMCG development cost and break-even models, July 2026. (2) NACS, "U.S. Convenience In-Store Sales Top $340 Billion," press release, April 15, 2026 (NACS State of the Industry, 2025 data). (3) NACS/NIQ TDLinx U.S. Convenience Store Count, released January 26, 2026; CSP Daily News coverage, April 2026. (4) NACS Magazine, "5 Key Metrics Defining the Convenience Industry's Health" and "In-Store Sales Drive Growth," June 2026; NACS "Who Makes Money Selling Gas?," February 2025. (5) U.S. Energy Information Administration, U.S. Product Supplied of Finished Motor Gasoline, annual series, release of May 29, 2026. (6) U.S. EIA, Short-Term Energy Outlook, April and July 2026 editions. (7) U.S. EIA, Annual Energy Outlook 2026, April 8, 2026; IEA World Energy Outlook 2025, November 12, 2025. (8) U.S. EPA, Semiannual Report of UST Performance Measures, End of FY2025 (data through September 30, 2025; published December 2025). (9) U.S. EPA, "UST and LUST Program Challenges in a Changing Transportation Sector," EPA 510-R-24-001, December 2024. (10) ASTSWMO, State Fund Survey Results, 2018 edition. (11) U.S. SBA, SOP 50 10 8, effective June 1, 2025 (technical updates effective March 1, 2026), including Appendices 6 through 8. (12) U.S. SBA, Policy Notice 5000-879058, Coordination of 7(a) and 504 Maximum Loan Limits, effective July 4, 2026. (13) U.S. SBA, Information Notice 5000-872051, FY2026 loan fees, August 28, 2025. (14) Yesway, Inc., Form S-1/A, April 13, 2026, SEC EDGAR; IPO pricing releases, April 2026. (15) Murphy USA Inc., FY2024 and FY2025 annual reports and quarterly earnings releases, SEC. (16) Casey's General Stores, Form 8-K, July 26, 2024 (Fikes/CEFCO), and FY2025-FY2026 earnings releases. (17) Alimentation Couche-Tard, press release, July 16, 2025; FY2025-FY2026 quarterly releases. (18) Sunoco LP, Form 8-K and 10-Q, Parkland Corporation acquisition, closed October 31, 2025. (19) Cumberland Farms Limited (EG Group), Form F-1, July 2, 2026, SEC EDGAR. (20) ARKO Corp. / ARKO Petroleum Corp., Form 8-K, February 2026 (IPO). (21) The Boulder Group, Net Lease Market Report and Net Lease Tenant Profiles, Q1 2026, released March 7, 2026. (22) U.S. EIA, Gasoline and Diesel Fuel Update: state motor fuel taxes as of January 1, 2026 (corrected March 4, 2026). (23) California State Water Resources Control Board, single-walled UST program updates, October 2025 and January 14, 2026; SB 445 (2014). (24) Phillips 66 announcement, October 2024 (Los Angeles refinery); Valero Energy notices and wind-down disclosures, April 2025 through Q1 2026 (Benicia). (25) California Legislative Analyst's Office cap-and-trade analysis; CARB LCFS amendment materials, 2024-2025; UC Davis ITS cost estimates. (26) City of Petaluma ordinance, February 22, 2021, and subsequent municipal ordinances via the CONGAS ordinance tracker; CSP Daily News. (27) FHWA, Traffic Volume Trends and Spring 2025 long-term VMT forecast. (28) U.S. EPA, Automotive Trends Report, 2025 edition (MY2024); S&P Global Mobility, average vehicle age release, June 27, 2025. (29) Cox Automotive / Kelley Blue Book, EV Market Monitor, October 2025, January 13, 2026, and March 25, 2026. (30) Public Law 119-21 (One Big Beautiful Bill Act), July 4, 2025; Congressional Review Act resolutions signed June 12, 2025 (ACC II waiver revocation). (31) sbalenders.com, "Understanding SBA Loan Failures," March 2025 (tabulation of SBA FOIA data, 1995-2024). (32) PeerSense, 2026 SBA Lending Report and industry data, updated May-July 2026. (33) LenderHawk, SBA Lenders for Gas Stations, 2026 (SBA data, FY2010-FY2025). (34) Getty Realty Corp., FY2025 Form 10-K and earnings materials, February 2026. (35) Seven & i Holdings, Q2 FY2024 presentation and October 2024 announcements; FY2026 store network plan; CSP Daily News and C-Store Dive coverage. (36) Philip Morris International, FY2025 results (Zyn shipment volumes). (37) NRC Realty & Capital Advisors, "Convenience Store Industry M&A Review and Outlook: 2025"; U.S. Bankruptcy Court, S.D. Tex., In re Mountain Express Oil, 2023. (38) Capstone Partners, Convenience Store and Fuel Distribution M&A Update, June 2025. (39) 40 CFR Part 280 (including Subpart H); 42 U.S.C. Section 9601(20)(E) (secured creditor exemption). (40) Federal Reserve Bank of Dallas, Working Paper 2509 (2025); Davis, McRae and Seira, Journal of International Economics (2025).
Comments