How Many Gas Stations Are in the US? Counts by State

The Census Bureau counted 109,174 gasoline stations with paid employees in March 2023, and 6,817 gasoline-station businesses without payroll. NACS, the convenience-store trade association, counted 120,061 fuel-selling stores at the end of 2023. The numbers differ because each source counts a different set of businesses.

The Census series is also shrinking. It showed 114,033 gasoline stations in 2012 and 109,174 in 2023, a drop of 4,859, while the number of people on those stations’ payrolls rose from 862,630 to 1,016,960. Fewer, larger stations is the national pattern: establishments with fewer than 10 employees fell 12.9% over the period, and establishments with 20 or more employees rose 81.3%. The state with the most stations is Texas, with 10,670. The state with the most per resident is Mississippi, with 64.4 per 100,000 people, three times the rate in California (20.7).

Three counts, three definitions

The sources in this article define a gas station differently, so each count starts from a different list. Table 1 sets out the figures that circulate, with the unit each one counts and the date it describes. The Census numbers come from its County Business Patterns (CBP) state files and its Nonemployer Statistics (NES) file for 2023. The NACS numbers come from its published store count, which it compiles with the NIQ TDLinx database and dates to December 31 of the year before the label on the page.

The smallest and most defensible count is the Census one, because its definition is written down and it can be rebuilt by anyone from the downloadable files. It leaves out every business that has no employee on the payroll in the week of March 12, and it classifies each business by what the business mainly does. The NACS counts are larger, and they are the better guide to how many places a driver can buy fuel from a store that also sells food. The 148,000 figure is the widest, and NACS describes it as an estimate, which is how it is used here.

What the Census Bureau counts, and what it leaves out

CBP covers establishments with paid employees. The Census methodology page defines an establishment as “a single physical location at which business is conducted,” measures employment in the week of March 12, and states that the series excludes self-employed individuals, together with businesses that have no employer identification number or no employees. A gasoline station run by its owner with no hired help is therefore outside CBP, however many gallons it pumps.

The industry definitions come from the 2017 NAICS, which the 2023 CBP file still uses. Within the Gasoline Stations subsector, NAICS 447110 is established for “retailing automotive fuels (e.g., diesel fuel, gasohol, gasoline) in combination with convenience store or food mart items,” in either a convenience-store or a gasoline-station setting. NAICS 447190 covers “gasoline stations (except those with convenience stores),” with truck stops and marine service stations as listed examples. The adjacent industry, NAICS 445120, covers “convenience stores or food marts (except those with fuel pumps).” A convenience store that sells fuel is outside 445120 by definition. That is the split that makes a like-for-like comparison with NACS possible later in this article.

The same Census workbook places warehouse clubs, superstores and supercenters in the general merchandise group, and each industry description is written for establishments “primarily engaged” in its activity. A fuel center counted as part of a supercenter establishment therefore falls outside NAICS 447. That is one reason the Census count can sit below the number of places where fuel is sold, and NACS’s list of fueling outlets carries 6,945 hypermarket and big-box sites on a line of its own.

For the businesses with no payroll, the Census Bureau runs Nonemployer Statistics, built from IRS business tax records. It describes a nonemployer business as one “with no paid employees, that are subject to federal income tax,” with receipts of $1,000 or more. The 2023 state file, coded in the 2022 NAICS, lists 6,817 businesses under 4571, Gasoline Stations. The 2022 NAICS groups gasoline stations (4571) with fuel dealers (4572) in one subsector, and 4571 is the code that corresponds to CBP’s 447. Those 6,817 businesses are 5.9% of the combined employer and nonemployer total of 115,991, and their average receipts were $394,000 each. Because the nonemployer file counts tax-filing businesses, this article calls them businesses, not stations. The nonemployer share varies by state: 10.7% of Mississippi’s combined 2,121 (226 businesses), 9.9% of Connecticut’s 1,177, and 9.5% of both Georgia’s 5,638 (536) and New Jersey’s 2,454 (232), against 2.0% in Maine, 2.1% in Idaho and Nevada, and 2.3% in Washington.

As of November 3, 2026, the Census directory of CBP datasets ends at 2023, so 2023 is the latest year for the employer counts. The nonemployer directory already lists 2024, which is why the 2023 nonemployer file is used here, to keep the two Census sources in the same year.

Rebuilding the count from the files

The CBP state files are comma-delimited text with one row per state, industry code and legal form of organization. To reproduce Table 1, keep the rows whose naics value is 447///, 447110 or 447190 and whose lfo value is a hyphen, which marks all forms of organization combined, then add the est column over the 51 state rows. The other lfo codes split the same establishments by ownership form, so adding them to the hyphen rows counts every establishment twice. The industry levels repeat as well: in 2023 the 4471// row carries the same figures as 447///, and 44711/ carries the same as 447110, so adding the levels together inflates the total. Column names change between years. The files for 2012 through 2016 label the smallest size class 1_4 and the files from 2017 on label it less than 5, and only the 2015 file prints its header in capitals.

What each count answers, and what it omits

Each count suits one question and misleads on the next. The Census employer series has a written industry definition and a twelve-year history on it, so it answers whether stations with payroll are becoming more or fewer and which states are dense with them. NACS’s store counts are the better guide to where a driver can buy fuel at a store that also sells food. The wider estimate of fueling outlets is the only denominator that matches an all-outlets gallons figure. Table 7 lines the five counts up against what each leaves out.

Raw counts and per-resident counts rank the states differently, so the choice matters for any comparison between places: Texas leads in raw counts at 10,670 but sits 28th per resident at 34.7, while Mississippi has 1,895 and leads per resident. For gallons per station, the denominator has to match the gallons. Dividing by Census stations only gives 3,303 gallons a day, against 2,436 on 148,000 outlets, a 36% difference between two defensible answers.

Fewer stations, more people on the payroll

The twelve CBP state files from 2012 to 2023 tell a story that a single-year count hides. Gasoline-station establishments peaked in this window in 2012 at 114,033, slipped to 111,076 in 2016, rose to 112,866 in 2017, and then declined to 109,174 by 2023. The 2023 count is 69 below 2022 and 3,692 (3.3%) below 2017. Employment moved the other way. Paid employees counted in the March week rose in nine of eleven year-to-year steps, from 862,630 to 1,016,960, and the average establishment went from 7.6 employees to 9.3.

The two industries moved differently. Gasoline stations with convenience stores fell 2.0% over the twelve years, from 97,939 to 96,002, and they are 87.9% of the 2023 total. The other gasoline stations, which include truck stops and stations without a store, fell 18.2%, from 16,094 to 13,172. Their average size went the opposite direction: 8.8 employees per establishment in 2012 and 12.8 in 2023, against 7.4 and 8.8 for stations with convenience stores. The stations left in that category are, on average, much larger sites.

The employee-size classes in Table 3 show where the shift happened. Establishments with 5 to 9 employees fell 18.6%, from 41,896 to 34,107. The smallest class, under 5 employees, fell 7.4%. Establishments with 10 to 19 employees rose 6.3%, those with 20 to 49 rose 70.4%, and the 50-or-more class went from 641 to 1,711, 2.67 times as many. The Census files do not say whether existing stations hired more people, whether small ones closed and larger ones opened, or both, and the counts alone cannot separate those. What they establish is the change in shape: three-quarters of stations had fewer than 10 employees in 2012 (74.5%), and two-thirds do now (67.8%).

Size differs by place as well as by year. Employees per establishment in 2023 range from 15.1 in Delaware, 14.1 in Minnesota and 13.8 in both Iowa and North Dakota down to 6.7 in Massachusetts, 6.4 in Georgia, 6.2 in Alabama, 5.6 in Connecticut and 5.5 in the District of Columbia. The industry’s annual payroll was $15.9 billion in the 2012 file and $29.7 billion in the 2023 file, up 86.7% against 17.9% more employees. Dividing annual payroll by the March headcount gives $18,435 per employee in 2012 and $29,193 in 2023, a nominal rise of 58.4%. Part-time schedules, turnover during the year and inflation all feed that ratio, so it indicates direction and is not a wage.

NACS’s own series looks different because it counts a different universe. Its all-store counts run from 149,220 (the count NACS labels 2013) to 154,958 (labelled 2018), 148,026 (labelled 2022) and 151,975 (labelled 2026), a spread of 4.7% between its lowest and highest labelled years, where the Census establishment count fell 4.3% over its own window. The two series cannot be spliced, and neither is a measurement error: CBP drops a business when its payroll disappears.

Stations per 100,000 people, state by state

The national average is 32.4 Census gasoline stations per 100,000 residents, or one for every 3,085 people. The median of the 51 places is 35.1. Seven states exceed 50 per 100,000: Mississippi (64.4), South Dakota (63.2), Alabama (60.4), Vermont (59.3), North Dakota (55.9), Wyoming (55.5) and Iowa (52.2). Twelve places fall below 25, from the District of Columbia at 11.8 and Hawaii at 16.5, through California (20.7), Arizona (21.3), Maryland (22.3) and New York and Oregon (both 22.4), to Washington, Delaware, Nevada, Utah and New Jersey. Among the 50 states, the highest rate is 3.9 times the lowest.

The raw counts follow population. Texas (10,670), California (8,116) and Florida (5,928) hold 22.6% of the national total between them, and the five largest, which add Georgia (5,102) and New York (4,419), hold 31.4%. At the other end the District of Columbia has 81 gasoline stations with payroll, Alaska 200, Hawaii 238 and Delaware 239.

A Mississippi resident has a gasoline station for every 1,553 residents. A California resident has one for every 4,830. The tables in this article cannot say why: they hold no data on land area, road miles, or how much of each state’s population is rural, and an explanation built on those would need those sources. They do show that the pattern is not a size effect of the stations. Mississippi’s 1,895 stations sell an average of 866,000 gallons a year, about 72% of the national mean, so the stations are many and each is smaller.

The 2012-to-2023 column adds a second layer. Forty-four of the 51 places lost stations. Vermont lost the largest share, 18.1%, ahead of Virginia (17.3%), Maine (16.4%) and Hawaii (15.9%), and Virginia lost the most in absolute terms, 591, followed by North Carolina (466), Texas (443) and Florida (424). Seven places gained: the District of Columbia (12.5%), Indiana (11.1%), California (8.0%), Tennessee (4.2%), Alabama (0.7%), Connecticut (0.4%) and Wisconsin (0.2%). California gained 603 and Indiana 288. The files do not report why a count changed, and a change in a state’s count can come from openings, closings, or an establishment moving between NAICS codes, which these tables cannot distinguish.

Gallons per station: a ratio that mostly measures the number of stations

EIA’s State Energy Data System estimates each state’s motor gasoline consumption. Its technical notes say the state values are built from sales data in the Federal Highway Administration’s Highway Statistics, scaled so that the states add up to the national product supplied total in EIA’s Petroleum Supply Annual, and that the national total includes the fuel ethanol blended into gasoline. For the transportation sector, the series used here (MGACP), SEDS counts highway gasoline plus boating and recreational-vehicle gasoline, less the special fuels that highway reporting mixes in. For 2023 that sums to 3,133,768 thousand barrels, or 131.6 billion gallons, and the 51 state values add exactly to the national row.

Divided by the 109,174 Census gasoline stations, 131.6 billion gallons is 1,205,582 gallons per station per year, or 3,303 gallons a day. That is a ceiling on what a typical station sells, not a measurement of it, for three reasons. The numerator includes gasoline sold at the outlets the Census count does not cover: the hypermarket and big-box sites and the kiosk sites in NACS’s list, and fuel sold through stores whose payroll sits under another code. It leaves out diesel entirely, which matters for any station whose main product is diesel. And it assigns gallons to the state where they were sold, as the tax reports do, not to the state where they were burned.

NACS publishes a number to test the arithmetic against: its page on who sells fuel says the average fueling location sells about 2,500 gallons a day, and puts the number of fueling outlets at about 148,000. Divide the same 131.6 billion gallons by 148,000 outlets and 365 days and the answer is 2,436. The two figures agree to within 3%, which is consistent with the gallons and puts the difference between 2,436 and 3,303 in the denominator. It is a check on the denominator, not independent proof of the gallons, because the same NACS page gives national fuel sales of 8.9 million barrels a day, about 370 million gallons, and 370 million divided by 148,000 outlets is 2,500. Counting only the employer gasoline stations raises the daily average by 36%.

The state column adds a second finding. Across the 50 states the rank correlation between stations per 100,000 people and gallons per station is -0.91: where stations are dense, each sells less. Gallons per station equals gallons per person divided by stations per person, and the second quantity varies more than the first. Per-person gasoline runs from 252 gallons (New York) to 599 (Alabama) among the 50 states, a factor of 2.4, while stations per person run through a factor of 3.9. The ratio therefore takes its ranking mostly from the station count. Delaware tops the gallons-per-station list at 2.01 million gallons a year, or 5,519 a day, and its gallons per resident, 465, are only 19% above the national 391. What sets it apart is that it has 23.1 stations per 100,000 residents, 43rd of 51, so each carries a lot. The reverse holds in Mississippi. Its residents burn 558 gallons each, 43% more than the national figure, and it still ranks 49th on gallons per station at 866,000, because it has more stations per person than any other state.

For a sense of scale in one state: California’s 8,116 stations and 12.47 billion gallons give 1,536,695 gallons per station, 4,210 a day. Texas’s 10,670 stations and 14.19 billion gallons give about 1.33 million, or 3,643 a day. Neither number is a typical station’s sales. Both would shrink if the denominator included the outlets outside the Census category. The deep dive on gasoline use per driver by state works from FHWA’s state gasoline tables, the sales data that SEDS is built on, with licensed drivers in the denominator, and shows how far a state ranking moves when the denominator changes. The lesson is the same here.

Bridging the Census count to the 120,061 stores NACS counts

Two like-for-like comparisons are possible, because the Census categories split along the same line as NACS’s: stores with fuel pumps against stores without. The first compares all convenience stores. NACS counted 152,396 at the end of 2023, and the Census counted 96,002 gasoline stations with convenience stores plus 36,992 convenience stores without fuel pumps in March 2023, or 132,994. NACS’s count is 19,402 higher, 14.6% above the Census sum.

The second compares stores that sell fuel. NACS counted 120,061 at the end of 2023. Table 5 steps through the Census categories from the narrowest to the widest and shows what is left unexplained at each step.

The gap shrinks from 24,059 to 4,070 as the Census categories widen. At that point the Census total, with the nonemployer businesses added, is within 3.4% of NACS’s fuel-selling count. The comparison has limits that the sources do not let anyone remove. The Census employer count is for March and the NACS count is for December. NACS’s fuel-selling count itself rose by 1,791 in 2024 and by 768 in 2025. NACS’s store count page does not state the definition of a convenience store that the count uses, so the stores in the 120,061 that a Census classification would place outside 447 cannot be identified. And 447190 includes truck stops and marine stations that a convenience-store list may or may not hold. The remaining 4,070 stores are therefore the sum of timing, classification and coverage differences, and no source here splits them.

A state-level check shows the two sources agree about geography even where their levels differ. NACS’s page lists its ten largest states by all convenience stores. In Table 6 each is set beside the Census sum of gasoline stations with convenience stores and convenience stores without pumps. The NACS counts are as of December 31, 2025 and the Census counts are for March 2023, so the dates are far apart, and the ratios are best read as bands.

The ratios run from 0.97 in New York to 1.23 in California and Michigan, with a combined 1.14. Of the ten, New York is the only state where the Census count is the higher of the two: it counts 4,264 convenience stores without fuel pumps against 3,560 gasoline stations with convenience stores, so more than half of its convenience stores in the Census files (54.5%) have no fuel pumps. Its density of gasoline stations is also among the lowest (22.4 per 100,000 residents). The ten states hold 52.1% of NACS’s national total and 52.3% of the Census sum, so the two lists put the stores in the same places even though the national levels are 14% apart.

NACS’s January 2026 release also splits its all-store count by owner size: companies with 10 or fewer stores own 95,672 stores, 63% of the total, and companies with 500 or more stores own 33,810, or 22.2%. A year earlier the figures were 95,946 (63.0%) and 34,042 (22.4%). The units differ from the Census size classes, which count employees at a site and not stores per owner, so the two do not reconcile. They agree in direction: most sites belong to small operators, and the Census has the small-site share (under 10 employees) at 67.8%.

The widest definition, NACS’s estimate of about 148,000 fueling outlets, rests on a NACS page dated June 19, 2026 that gives its breakdown without a date: more than 120,000 convenience stores that sell fuel, 12,659 kiosk fueling sites from NIQ, 6,945 hypermarket and big-box sites from Energy Analysts International, and roughly 6,000 service stations that NACS labels an estimate. The components come from different providers and may describe different years, so they are not summed here, and the 148,000 is used only to test gallons per day.

One state, three denominators

Take Mississippi. The Census counts 1,895 gasoline stations with payroll, 1,760 with convenience stores and 135 without. The state had 2,943,172 residents on July 1, 2023, so 1,895 divided by 2,943,172 and multiplied by 100,000 is 64.4 stations per 100,000 people, or one per 1,553 residents. EIA’s SEDS puts Mississippi’s 2023 transportation gasoline at 39,075 thousand barrels, which is 1,641.2 million gallons. That divided by 1,895 stations is 866,042 gallons per station a year, 2,373 a day, against the national 3,303. Add the 226 nonemployer gasoline-station businesses in the NES file and the count of businesses becomes 2,121, which lowers the per-business figure to 773,762 gallons. The number moved 10.7% without a single new pump, which is the sense in which gallons per station is a statement about the denominator.

For the price a station posts, the taxes and blend rules behind it are in why gas prices differ by state, and the slice a station shares with its supplier is in the gasoline price breakdown. Diesel, which the gallons here leave out, is covered in the diesel price guide, and the weekly prices posted at stations are on the PetroEyes gas price tracker. A parallel counting problem for the supply side of the same industry, with a census that defines its unit and a count that moves with the definition, is set out in how many oil wells are in the US.

Data and sources, all checked November 3, 2026: U.S. Census Bureau, County Business Patterns methodology and the 2023 state file (with the 2012 through 2022 state files in the same directory tree, for example 2012); the Nonemployer Statistics 2023 state file; the Census Bureau’s 2017 NAICS descriptions and 2022 NAICS descriptions; and the Vintage 2024 state population estimates (July 1, 2023 column). U.S. Energy Information Administration, State Energy Data System: consumption in physical units (series MGACP, 2023) and the petroleum consumption technical notes. NACS: Industry store count (2026 release, last updated November 3, 2026), January 26, 2026 release, January 31, 2025 release, January 23, 2024 release and Who Sells America’s Fuel. Puerto Rico and the island areas are not in the tables. Ratios, rankings, changes and gallons per station are computed from the sources listed.