Gas Station Feasibility and EV Exposure Across a 25-Year Term

The loan outlives the forecast

This is the structural problem with the asset class, and almost every study handles it badly. The two common failures are opposite and equally useless. One is silence, a study that projects fuel volume flat for ten years and never mentions electrification, which any reviewer will notice. The other is theatre, a study that quotes a long range projection to two decimal places as though the number were a measurement.

Three things are genuinely knowable over a 25-year horizon, and together they are enough to underwrite on.

The first is the arithmetic of fleet turnover, which is measured rather than modelled and which governs how fast anything can change. The second is how much of the site's gross profit already comes from something other than fuel, which on the industry numbers is a great deal more than most people assume. The third is the schedule of capital events on the physical asset, particularly the tanks, which is knowable to the year and which lands squarely inside the loan term.

Everything else is a scenario, and should be presented as one.

What is measured

United States motor gasoline consumption averaged 8.9 million barrels per day in 2025, 1 per cent below 2024 and 4 per cent below pre-pandemic demand in 2019. The same figure held in 2023 and 2024. The 2016 to 2019 average was 9.3 million barrels per day, and the record annual level, about 392 million gallons per day, was set in 2018. The 2019 peak has not been exceeded in any year since.

A technical note that matters for how the number is quoted: the federal agency does not directly survey petroleum consumption. It calculates product supplied, which is its proxy. Anyone citing gasoline consumption to a lender is citing a derived series, and should say so.

The single most important measured fact for a fuel site is not the level. It is the decoupling. Vehicle miles travelled reached an all-time high of 9.0 billion miles per day in 2024, above the 2016 to 2019 average of 8.8 billion, and gasoline consumption still fell, because increasing fuel efficiency and electrification of the fleet reduced gasoline use despite the increased driving.

Read that against how site level feasibility is normally argued. Traffic count is the standard demand proxy for a fuel site, and traffic counts can rise while the gallons sold across the same forecourt fall. A study that projects volume from traffic growth alone has built in an error that the national data has already demonstrated. Volume per vehicle is falling, and it is falling for reasons that have nothing to do with the site.

What is projected, and why it is not a forecast

The long range outlook that everyone cites carries a disclaimer that almost nobody quotes. The agency describes its outlook as best understood as a product suite for alternative futures analysis, not a set of predictions, and states that the results are not predictions of what will happen but modelled projections of what could happen given certain assumptions and methodologies. Of its own reference case it says plainly that it should not be regarded as the most likely of the cases.

There is a second problem with the reference case, and it is specific to this moment. The 2025 edition, published on 15 April 2025, assumed that laws and regulations in effect as of December 2024 would remain in effect through 2050, and stated explicitly that legislation, regulations, executive actions and court rulings after that date were not included. A great deal happened after that date, all of it in the same direction, and the effect is that the reference case now assumes emissions and fuel economy rules that no longer exist.

The direction of travel across editions is also worth knowing. The 2020 edition projected that motor gasoline declines by 19 per cent through 2050. Independent analysis of the 2025 edition's data tables puts the reference case decline near 45 per cent by 2050, roughly double, attributed to faster projected adoption and to the vehicle emissions standards adopted in April 2024. That 45 per cent figure is an analyst's derivation from the published tables rather than a sentence the agency wrote, and should be presented that way.

The practical consequence for a study is a defensible scenario band rather than a point estimate. The reference case, built on assumed aggressive electrification, is the low volume bound. The alternative transportation case, which assumes the emissions and fuel economy rules are not in place and is therefore the case closest to current law, is the high volume bound. The distance between them is the honest uncertainty, and presenting it is more credible than resolving it.

Three electric vehicle numbers, and only one of them matters here

The figure quoted at lenders is almost always the wrong one, because the three available measures diverge by a factor of four.

Share of new sales. Battery electric vehicles were about 7.5 per cent of new light duty sales across 2025, with plug in hybrids at about 1.6 per cent. In the second quarter of 2026 battery electric was 6 per cent of new sales, down from 7 per cent a year earlier, and plug in hybrids fell from 1.9 per cent to 1.4 per cent. Conventional hybrids, which never plug in and burn only liquid fuel, hit a record 16 per cent of 2025 sales.

Share of vehicles in operation. In 2024, the most recent data year, electric vehicles accounted for 2 per cent of all registered light duty vehicles in the United States. An industry association put the figure at 2.5 per cent as of the fourth quarter of 2025.

Share of miles driven. Not published as an official federal series. Any figure in circulation is modelled.

The gap between roughly 9 per cent of sales and roughly 2 per cent of the fleet is the whole story, and it is not a rounding artefact. Annual sales are small against the installed base. A site sells fuel to the fleet, not to the showroom.

One more measured point, because it cuts against the trend line most studies assume. Battery electric sales reached a record 12 per cent share in September 2025, immediately before the federal purchase credits expired, then fell below 6 per cent for the remaining months of the year. 2025 was the first year in which annual battery electric sales and market share both declined.

The arithmetic that actually governs

Here is the calculation that belongs in every fuel site study, because it is built entirely from measured data and it constrains every scenario above.

The average age of a United States light vehicle reached 12.8 years, a record high, in the 2024 data, up two months for the second consecutive year. There are 289 million vehicles in operation, retiring at a steady scrappage rate of 4.5 per cent a year. New registrations passed 16 million in 2024 for the first time since 2019, which was still not enough to offset the growing volume of aged vehicles in the fleet.

Put those together. At a 4.5 per cent annual scrappage rate against an installed base of 289 million, and with battery electric vehicles at roughly 7.5 per cent of sales and falling, the operating fleet stays overwhelmingly liquid fuelled deep into the 2040s under any plausible sales trajectory. Even a dramatic reversal in the sales mix takes more than a decade to reach the pumps, because the mechanism by which a sales change becomes a fuel demand change is the scrapping of an old vehicle, one at a time, at 4.5 per cent a year.

This is the defensible foundation for a 20 to 25 year volume assumption, and its strength is that it is arithmetic rather than opinion. It does not require the study to take a view on adoption. It only requires the reviewer to accept that a car sold in 2026 is still on the road in 2038.

Where the site is matters more than where the country is

National adoption figures are close to useless at site level, because the distribution is extreme.

California holds roughly 35 per cent of all United States electric vehicle registrations, about 1.26 million light duty vehicles in the 2023 data, and leads on every measure, at roughly 4.1 per cent of its registered fleet. The next highest jurisdictions in 2023 were the District of Columbia at 3.5 per cent and Hawaii at 2.8 per cent. At the other end, North Dakota had 959 electric vehicles registered in the whole state in 2023.

The rural picture is starker still. Federal transport analysis finds the rate of adoption remains 40 per cent lower in rural areas. Research reporting on county level registration densities finds non metro counties often showing fewer than five, and in many cases zero, registered electric vehicles per 10,000 people, against 10 to more than 100 per 10,000 in metropolitan counties. A Harvard led study identified 34 counties that previously had public charging and became charging deserts, with another 36 losing most of their stations. A 2025 survey of rural Michigan drivers found 5 per cent would choose a battery electric vehicle as their next purchase.

For a study, the implication is direct. Electrification exposure is a trade area variable, not a national one, and the trade area figure is available from state registration data. A rural interchange site in a low adoption state and an urban site in California carry exposures that differ by an order of magnitude, and a study that applies the national number to both has told the lender nothing.

The policy reversal, and why it belongs in a scenario rather than a base case

Between mid 2025 and early 2026 the federal framework that earlier projections assumed was dismantled, and the dates matter because they explain why any pre 2025 forecast reads differently now.

Federal purchase credits for new, used and commercial clean vehicles terminated for vehicles acquired after 30 September 2025, under legislation signed on 4 July 2025, accelerating the previously scheduled sunset by more than seven years. Sales responded immediately and visibly, as described above.

The Environmental Protection Agency published a final rule on 18 February 2026 rescinding the 2009 endangerment finding and repealing all greenhouse gas emission standards for light, medium and heavy duty motor vehicles and engines, both prospectively and for prior model years, with an effective date of 26 April 2026. Petitions for review have been filed, and the outcome is not settled.

The maximum corporate average fuel economy civil penalty was reset to zero by the same July 2025 legislation, and the regulator issued an interpretive rule on 11 June 2025 concluding it is improper to consider electric vehicle fuel economy when setting standards, followed by a proposed rule on 5 December 2025 to reset standards on the basis of petrol and diesel vehicle performance only.

Three congressional resolutions signed on 12 June 2025 nullified California's vehicle emissions waivers. Eleven states sued the same day, and the litigation is unresolved.

Federal corridor charging funding was suspended by a memorandum of 6 February 2025, partially unfrozen by a preliminary injunction in June 2025, and restored on 23 January 2026 when a court found the suspension arbitrary, capricious and in excess of statutory authority and vacated the plan revocations. Separately, a January 2026 rescission clawed back almost $504 million, an estimated 3,700 charging ports, from 30 states that had not obligated all first year funding.

A feasibility study should record all of this as dated fact and then decline to extrapolate from it. Policy that reversed this fast in one direction can reverse again, and much of it is under active challenge. Its correct place in a study is as the input that defines the high volume scenario, not as the basis for a new base case.

The profit was never mainly fuel

The most useful reframing available to a fuel site study is that the asset has been converting itself out of fuel dependence for two decades, and the industry data shows it plainly.

Convenience store in store sales reached a record $341.2 billion in 2025, up 1.7 per cent, the twenty third consecutive record year. Total industry sales, fuel and in store combined, were $817.5 billion, of which fuel accounted for $476.3 billion, down 5.4 per cent on a lower average pump price of $3.11 against $3.30, even though gallons sold rose 0.5 per cent.

That last sentence is worth reading twice. Fuel revenue fell by more than 5 per cent in a year when volume rose, because fuel revenue tracks price. Any study that models fuel dollars rather than fuel gallons and margin has built a revenue line that moves with a commodity price the operator does not control.

Foodservice accounted for 28.5 per cent of in store sales in 2025 but 38.9 per cent of in store gross profit dollars, up from 27.7 per cent and 38.6 per cent in 2024. Prepared food is now 73.9 per cent of foodservice sales, up from 66.4 per cent in 2021. For scale of the shift, foodservice was 11.9 per cent of in store sales in 2005.

Fuel margins, where they can be measured, come from public operators rather than an industry average: one reported retail fuel margin of 44.9 cents per gallon in the second quarter of 2025 and 43.6 cents in the third, another about 40.2 cents. These are company specific figures before card fees, and card fees are not trivial: the industry paid a record $21.3 billion in credit and debit card fees in 2025, with direct store operating expenses up 4.2 per cent.

The structural picture is a store with a fuel forecourt attached, where fuel is a high volume, thin and volatile margin traffic generator and the profit is made converting the visit. The store count reflects it: 151,975 convenience stores as at 31 December 2025, down 280 or 0.2 per cent, the second consecutive annual decline, while the number of stores selling fuel rose by 768 to 122,620, the highest in eight years. Sites are not disappearing. The weak ones are, and the survivors sell fuel.

For a feasibility study the consequence is a change in what gets forecast. Gallons and cents per gallon are one line. Inside sales and their gross profit are another, with foodservice capability treated as a physical and operational attribute of the site rather than an afterthought. A site with a kitchen, a drive through or a car wash carries a materially different exposure to fuel volume decline than a site with a cooler and a coffee pot, and that difference is visible in the financial model or it is not in the study at all.

Charging is not the hedge it is sold as

The instinct to answer electrification exposure by putting chargers on the forecourt runs into measured economics that do not yet support it as a revenue line.

The strongest available evidence is the audited performance of the largest pure play network. Across 2025 it reported revenue of $384 million, up 50 per cent, with charging network revenue of $218 million and network throughput of 366 gigawatt hours, and achieved its first ever positive full year adjusted earnings before interest, tax, depreciation and amortisation of $12.0 million. In the same year it reported a net loss of $95.4 million. Its network utilisation reached about 24 per cent in the fourth quarter, against roughly 19 per cent a year earlier, and its average daily throughput per stall was running well below its own three to five year target.

Against that, the national picture. Independent tracking put national fast charging utilisation at 15.6 per cent in the first quarter of 2026, down from 16.2 per cent a year earlier, and analysis of the sector places modelled breakeven near 15 per cent. The median public fast charging site sits at the breakeven line rather than comfortably above it.

The cost side explains why. Federal laboratory data places direct current fast chargers at up to $150,000 per unit installed, with independent work citing roughly $90,000 to $170,000 per port fully installed including civil works and grid connection. Electricity cost for fast charging has been found to range from under $0.10 to over $2 per kilowatt hour depending on the tariff, with low utilisation the dominant driver, and peer reviewed corridor analysis finds that stations subject to demand charges cost around 40 per cent more, with a modelled twenty year breakeven averaging $0.42 per kilowatt hour and fewer than half of stations reaching cost parity with petrol on a per mile basis.

None of this argues against installing chargers. It argues against underwriting them. On current evidence, charging belongs in a study as an amenity with disclosed utilisation and demand charge sensitivities, not as a revenue line carrying coverage. Vendor payback claims should be labelled as vendor claims.

The capital events are schedulable, and they land inside the term

This is the part of the exposure that is not speculative at all, and it is routinely underweighted.

The physical benchmark is the tank warranty, which is typically 30 years, structured as ten years plus a further twenty. The 30-year mark is widely treated in the industry as a tank's expiry, insurers apply a marked premium increase at that point, and more than half of the roughly 415,000 tanks in service have passed 25 years. Reporting of regulatory and state association research indicates around 40 per cent of operating tanks are more than 30 years old with an average age of 25 years, while the environmental regulator's own estimate of average historical tank lifespan, based on age at removal, is closer to 20 years. Actual longevity varies widely, and tanks operating past 40 years exist.

Set that against a 25-year loan. A site with 20-year-old tanks at closing will face a tank replacement decision inside the term, probably in the middle third of it, and the study should say so with a year attached rather than leave the reserve line unexplained.

The compliance schedule is equally concrete, and it is federal. Under the 2015 revision of the underground storage tank regulations, the first major revision since 1988, initial testing of spill prevention equipment, containment sumps and overfill prevention had to be completed by 13 October 2018, with overfill prevention inspected at least every three years and spill prevention equipment and sumps tested at least every three years thereafter. Walkthrough inspections became mandatory: every 30 days for spill prevention equipment and release detection equipment, and annually for containment sumps and hand held release detection equipment, with records retained for one year. Operator training obligations attach to three classes of operator, with Class A and B operators designated after October 2018 required to be trained within 30 days of assuming duties and Class C operators before assuming them. Financial responsibility runs to $1 million per occurrence for petroleum marketing facilities or sites averaging more than 10,000 gallons a month, with an annual aggregate of $1 million for one to 100 tanks and $2 million above that, excluding legal defence costs and without limiting liability. Tanks and piping installed or replaced after 11 April 2016 must be secondarily contained with interstitial monitoring, and new dispenser systems installed after that date require under dispenser containment. States with approved programmes may be more stringent, and many are.

On reserves, there is no published standard to lean on. No agency, supervisory body or appraisal authority publishes a replacement reserve figure specific to fuel retail. The concept is generic: an annualised allowance for the eventual replacement of short lived components that wear out before the building does. That means the reserve has to be built from a property condition assessment of the actual site, component by component, with tank age as the largest single input, rather than taken from a rule of thumb. A study that carries a percentage of revenue as a reserve, with no component schedule behind it, has not addressed the question.

The term structure problem nobody names

There is a structural mismatch inside the loan itself, and it deserves to be visible in the study.

SBA real estate maturities run to a maximum of 25 years including extensions, with an additional period available to complete construction. Equipment maturities are ten years or less, unless the loan finances assets with a useful life exceeding ten years, and machinery and equipment maturities cannot exceed the average life expectancy of the equipment. Mixed use of proceeds produces a blended, weighted average maturity. On the 504 side, debentures are available in 10, 20 and 25 year maturities, with the 25 year debenture available for projects approved on or after 2 April 2018 and not available by modification to projects approved before that date.

Now compare the asset. For tax purposes the fuel marketing equipment, meaning tanks, pipelines, dispensers and the canopy itself, sits in a five year recovery class, while the building and land improvements, including the canopy's concrete footings, sit in a fifteen year class. Those are tax recovery periods rather than engineering lives, but they are the closest published component life figures available, and the direction they point is unambiguous.

The practice the appraisal literature describes is that permanently attached fuelling improvements are commonly treated as real property, which allows them to be blended into the longer real estate amortisation. The economics of that are worth stating plainly for the lender: a 25 year amortisation is being applied to components with a service life a fraction of that length, which is precisely why the replacement reserve carries so much weight on this asset and why the appraisal's allocation between land, building, equipment and intangibles is worth reading closely rather than accepting.

That allocation is not optional on an SBA file. Gas stations, with or without a convenience store, sit on the special purpose property list, which requires an independent appraisal by a Certified General Real Property Appraiser who has completed no fewer than four going concern appraisals of equivalent special use property within the last 36 months, allocating separate values to land, building, equipment and intangible assets.

The environmental file, and the one case that is simply ineligible

Fuel sites carry their own appendix in the SBA rulebook, covering gas station environmental requirements, and the framework carries forward into the edition effective 1 October 2026 without substantive change, per published analysis of that edition.

The mechanics are tiered. A loan up to and including $250,000 on a property outside an environmentally sensitive industry requires only an environmental questionnaire. Above $250,000 a records search with risk assessment is added. Where the questionnaire or the risk assessment indicates elevated or high risk, or where the industry code matches the environmentally sensitive list, a Phase I environmental site assessment is required. Gas stations have tanks and match the sensitive code list, so the Phase I is not a judgement call. Where a Phase I identifies recognised environmental conditions, a Phase II follows.

Two timing points matter for scheduling. Environmental reports must be dated within one year of the date the SBA loan number is issued, which measures from the wrong end of the process for anyone planning backwards from submission. And the reliance letter accompanying the report is a mandatory form, under which the environmental professional certifies errors and omissions cover of at least $1 million per claim, waives dollar limitations on liability up to $1 million and time limitations other than statutes of limitation, waives indemnification from the lender and SBA, and certifies independence from the seller, borrower and operating company.

Where contamination is confirmed, the recognised routes to proceed are an indemnification agreement executed by a person with sufficient financial resources to cover remediation, an escrow account covering an approved remediation plan, a solvent government cleanup fund that has unconditionally agreed to cover the cost, or lender liability environmental insurance. Reliance on a state trust fund therefore requires that the fund be both solvent and unconditionally committed, which is a question with a different answer in different states and in different fiscal years.

One case is ineligible outright rather than expensive: where a third party will not waive its environmental indemnification right. That is a deal killer to check early rather than late.

How to write the exposure into the study

Nine things, all of them derived from the material above.

Forecast gallons and cents per gallon separately from fuel dollars, because fuel revenue moves with pump price and the operator does not control it.

Model inside sales and their gross profit as a separate line, with foodservice capability treated as a physical attribute of the site. The industry earns 38.9 per cent of in store gross profit from a category that is 28.5 per cent of in store sales, and a site that cannot participate in it is exposed differently.

Anchor the volume base case on fleet turnover arithmetic rather than on any published long range projection, and show the arithmetic. It is measured, it is simple, and it survives review.

Present the long range projections as a band with both bounds named, the reference case as the low volume bound and the case assuming no emissions and fuel economy standards as the high volume bound, and state which assumptions each rests on and as of what date.

Use trade area adoption data, from state registration records, not national figures. The spread between states is more than fourfold, and rural adoption runs 40 per cent below the national rate.

Do not project traffic growth into volume growth. The national data already shows record miles driven alongside falling gasoline consumption.

Put the capital events on a dated schedule inside the loan term, with tank age as the anchor, and build the replacement reserve from a component schedule rather than a percentage, because no published fuel retail reserve standard exists to cite.

Treat charging as an amenity with disclosed utilisation and demand charge sensitivities, and label vendor payback claims as vendor claims.

Check the local rules before writing anything about exit value. Several California municipalities have enacted permanent bans or moratoriums on new gas stations, one of which also prohibits adding pumps at existing sites, and at least two jurisdictions outside California have limited new stations or imposed a moratorium. Those are enacted. Other proposals, including a general plan provision that would ban new traditional stations and permit upgrades only for charging, are proposals. The difference matters, and only the enacted ones belong in a value conclusion.

Frequently asked questions

Is gasoline demand actually falling?

Slowly, and from a plateau. Consumption averaged 8.9 million barrels per day in 2025, 1 per cent below 2024 and 4 per cent below 2019, and the 2019 pre-pandemic level has not been regained. The more useful fact for a site is that vehicle miles travelled hit an all-time high in 2024 while gasoline consumption fell, because efficiency and electrification reduced fuel use per mile. Traffic and volume have separated.

Can anyone forecast fuel volume 25 years out?

No, and the agency publishing the best known long range outlook says so, describing its work as alternative futures analysis rather than a set of predictions and noting that its reference case should not be regarded as the most likely case. A study should present a scenario band with both bounds and their assumptions, not a single projected line.

What share of vehicles is actually electric?

About 2 per cent of registered light duty vehicles in the 2024 data, against roughly 7.5 per cent of new sales in 2025. The gap is the point: sales are small against an installed base of 289 million vehicles. Share of miles driven by powertrain is not published as an official series, so any such figure is modelled.

How fast could electrification actually reach the pumps?

Slowly, by arithmetic. The average light vehicle is 12.8 years old, a record, and the fleet retires at about 4.5 per cent a year. A change in the sales mix only becomes a change in fuel demand as old vehicles are scrapped, one at a time. That holds the fleet overwhelmingly liquid fuelled deep into the 2040s under any plausible sales path.

Does the 2025 federal policy reversal change the outlook?

It changes the inputs of every forecast written before it. Purchase credits ended for vehicles acquired after 30 September 2025, vehicle greenhouse gas standards were repealed in a rule published in February 2026, fuel economy penalties were set to zero and California's waivers were nullified. Battery electric sales fell in 2025 for the first time. Much of it is under legal challenge, which is why it belongs in a study as a scenario input rather than as a new base case.

Should a fuel site install electric vehicle charging?

Possibly, but not as a revenue line. The largest pure play network reached positive adjusted earnings before interest, tax, depreciation and amortisation for the first time in 2025 while still posting a net loss of $95.4 million, national fast charging utilisation sits near the modelled breakeven, installed cost runs to six figures per port, and electricity cost varies enormously with the tariff and utilisation. Underwrite it with disclosed sensitivities or leave it out of coverage.

What is the biggest underwriting risk on a fuel site over 25 years?

On the evidence, not electrification. It is margin and capital events: fuel gross margin net of card fees, which reached a record $21.3 billion industry wide in 2025, and the replacement schedule on tanks and forecourt equipment, which lands inside the loan term and has no published reserve standard behind it.

How should the replacement reserve be set?

From a component schedule built on the actual site, with tank age as the anchor. The common warranty benchmark is 30 years, insurers step premiums up at that point, and around 40 per cent of operating tanks are already past it. No agency or appraisal body publishes a fuel retail specific reserve figure, so a percentage of revenue with nothing behind it will not survive a careful review.

Sources

1. US Energy Information Administration, US Product Supplied of Finished Motor Gasoline, weekly and monthly series
2. US Energy Information Administration, Frequently Asked Questions, gasoline consumption, updated 29 March 2024
3. US Energy Information Administration, Today in Energy, transportation fuel demand and vehicle miles travelled, 2025 and 2026
4. US Energy Information Administration, Today in Energy, understanding petroleum product supplied, 19 September 2024
5. US Energy Information Administration, Annual Energy Outlook 2025, released 15 April 2025, and accompanying Today in Energy
6. US Energy Information Administration, Annual Energy Outlook 2020, transportation projections
7. US Energy Information Administration, Annual Energy Outlook 2026, released 8 April 2026
8. Labyrinth Consulting Services, analysis of Annual Energy Outlook 2025 data tables, 29 April 2025
9. US Energy Information Administration, Today in Energy, electric and hybrid vehicle sales, 9 February 2026
10. Alliance for Automotive Innovation, Get Connected Electric Vehicle Report, 2025 quarterly editions
11. US Department of Energy, Alternative Fuels Data Center, electric vehicle registrations by state
12. Argonne National Laboratory, plug-in electric vehicle registration share by state
13. Environmental and Energy Study Institute, rural electric vehicle adoption analysis
14. US Department of Transportation, rural electric vehicle adoption rate, cited in peer-reviewed literature
15. Scientific Reports, barriers to electric vehicle adoption in rural Michigan, 2025
16. Public Law 119-21, signed 4 July 2025, clean vehicle credit termination
17. Internal Revenue Service, FS-2025-05, 21 August 2025
18. Federal Register, rescission of the greenhouse gas endangerment finding and motor vehicle greenhouse gas standards, 91 FR 7686, 18 February 2026
19. Federal Register, National Highway Traffic Safety Administration, Resetting the Corporate Average Fuel Economy Program, 90 FR 24518, 11 June 2025
20. National Highway Traffic Safety Administration, proposed corporate average fuel economy reset rule, 5 December 2025
21. Congressional Review Act resolutions on California vehicle emissions waivers, signed 12 June 2025
22. Washington v. United States Department of Transportation, Western District of Washington, summary judgment 23 January 2026
23. 23 CFR Part 680, National Electric Vehicle Infrastructure Formula Program minimum standards
24. S&P Global Mobility, average age of United States light vehicles, released 21 May 2025
25. NACS and NIQ TDLinx, Convenience Industry Store Count, 2025 and 2026 editions
26. NACS, State of the Industry data, 2024 and 2025
27. ARKO Corp and Casey's General Stores, quarterly results filings, 2025
28. EVgo Inc, third quarter 2025 and full year 2025 results
29. National Renewable Energy Laboratory, electricity cost for electric vehicle fast charging, and the 2030 National Charging Network
30. McKinsey and Company, public electric vehicle fast-charging station profitability analysis
31. Paren, national direct current fast charging utilisation tracking, first quarter 2026
32. ScienceDirect, economics of electric vehicle corridor fast charging in the United States
33. SBA SOP 50 10 8, Lender and Development Company Loan Programs, effective 1 June 2025, including Appendix 5, Appendix 6, Appendix 7 and Appendix 8
34. SBA Information Notice 5000-880695, Issuance of SOP 50 10 8.1, 14 August 2026
35. SBA SOP 50 57, 7(a) Loan Servicing and Liquidation
36. 13 CFR Part 120, including Section 120.921
37. Federal Register, 504 Loans and Debentures With 25 Year Maturity, 83 FR 14536, 4 April 2018
38. 40 CFR Part 280, technical standards for owners and operators of underground storage tanks, including Sections 280.20, 280.35, 280.36, 280.93 and Subpart J
39. Federal Register, Revising Underground Storage Tank Regulations, 80 FR 41566, 15 July 2015
40. US Environmental Protection Agency, Learn About Underground Storage Tanks, and secondary containment guidance under the Energy Policy Act of 2005
41. Energy Policy Act of 2005, Public Law 109-58, Title XV, Subtitle B
42. Steel Tank Institute, STI-P3 warranty and SP001 inspection standard
43. Internal Revenue Service, Revenue Procedure 87-56, Asset Classes 57.0 and 57.1, and Internal Revenue Code Section 168(e)(3)(E)(iii)
44. Appraisal Institute, The Appraisal of Real Estate and The Dictionary of Real Estate Appraisal, reserve for replacement
45. Journal of Contaminant Hydrology, fuel release at retail service stations
46. Petaluma, California, ordinance prohibiting new gas stations, adopted March 2021
47. California Green Building Standards Code, 2019, 2022 and 2025 editions
48. North American Industry Classification System, 2022 revision, Subsector 457