Restaurant Software in the US 2026: The Independent Owner Is the Buyer, Payments Pay the Bills

Restaurant software looks like a crowded, solved market. Toast is public, point-of-sale terminals are in almost every fast-food counter, and every delivery app has a tablet on the pass. If you are thinking about building for US restaurants, the useful question is not "is there demand?" but "who exactly signs the contract, how much margin do they have to pay me, and what does one location have to be worth for my sales cost to make sense?"

Government statistics and the filings of listed vendors answer most of that. The short version: the buyer is usually a small, single-location owner; that owner keeps a few cents of every dollar; and the vendors that win make most of their money on payments, not on software.

The market: large, growing, and made of small firms

US food services and drinking places sold $1,032.7 billion in 2025, up from $984.2 billion in 2024 (US Census Bureau, Monthly Retail Trade). Growth has not stopped: September 2026 sales were 7.0% above September 2025, and January–September 2026 was 4.4% ahead of the same period a year earlier (Census, Advance Monthly Retail Sales). Part of that is price — full-service menu prices rose 3.5% in the twelve months to August 2026 (BLS, CPI).

Those sales run through 721,796 establishments, of which 264,800 are full-service restaurants, 269,945 limited-service restaurants, 87,158 coffee, juice and snack bars, and 42,249 bars (BLS, QCEW, 2025).

The more important number for a software founder is how those locations are owned. In 2022, 85.9% of full-service restaurants belonged to single-location firms, and those firms earned 69.2% of full-service revenue (Census, Economic Census 2022). The market is not concentrated: the four largest full-service firms took 6% of revenue, the fifty largest 15%. Limited-service is different — only 56.8% of locations are single-unit, earning 39.5% of revenue — and the latest franchise table, from 2017, put 57.3% of limited-service locations under a franchise.

Size by staff confirms the picture. Of 224,680 full-service employer firms, 159,252 had fewer than 20 employees (Census, Statistics of US Businesses, 2022) — about 71% by our calculation. By revenue, the densest band is $1M–$2.5M a year with 59,479 firms, followed by $500,000–$999,999 with 47,623 and $250,000–$499,999 with 28,808 (Economic Census 2022). Together that is about 136,000 firms between a quarter-million and two-and-a-half million dollars in sales — our sum of the three bands.

That is your realistic buyer if you sell to full-service restaurants: an owner-operator with one location and around a million dollars in sales, who decides alone and quickly, and churns just as quickly. Selling to fast food is a different motion — a franchisee who often has to use what the brand chooses, or a chain head office with procurement.

What the buyer can afford

Restaurant margins are thin at every level you can measure. Among 33 US-listed restaurant companies with at least $50M in revenue, the median operating margin in 2025 was 4.3%, the 25th percentile 0.2%, and 21.2% made an operating loss (SEC EDGAR financial data). The National Restaurant Association's operator survey — a proxy, mostly independents — put median pre-tax income at 2.8% of sales for full-service and 4.0% for limited-service operators in 2024. Only 56.6% of food service employer firms reported a profit in 2022 (Census, Annual Business Survey).

Restaurant-level margins in chain filings look healthier — 25.4% at Chipotle and 12.6% at Noodles & Company in 2025 — but they are company-defined and measured before corporate overhead, so they show what a well-run unit earns, not what is left for the owner.

Labour is the line that decides everything. Salaries, wages and benefits took a median 36.5% of sales at full-service operators and 31.7% at limited-service operators in 2024 (National Restaurant Association, proxy); the median labour expense line among listed chains was 31.2% of revenue in 2025 (SEC EDGAR). Average annual pay is $31,394 per full-service employee and $24,217 per limited-service employee, part-time staff included (BLS, QCEW, 2025), and the quits rate in accommodation and food services averaged 4.24% a month in 2025 (BLS, JOLTS) — summed over a year, about 51 quits per 100 jobs by our calculation.

Put those together for one restaurant. At a 2.8% pre-tax margin, a full-service restaurant with $1.19M in card sales keeps about $33,000 a year (our calculation). Any product priced as a cost line competes for a share of that $33,000. A product that saves labour hours or reduces turnover competes for a share of a labour bill that, at 36.5% of the same sales, is about $434,000 — more than ten times larger (our calculation). That is the difference between a nice-to-have and a budget line.

What one location is worth to a vendor

Toast is the cleanest public benchmark for a US restaurant platform. It had about 180,000 locations in June 2026 and an annualised recurring run-rate of $2.4B, growing 25% year over year (Toast, SEC filings). Recurring revenue per location was about $13,300 a year — but only $5,707 of that was subscription software in 2025. The rest is payments: Toast processed about $1.19M a year per location, kept 2.58% of payment volume as revenue, and 0.59% of volume after paying processing costs (Toast, SEC filings, 2025).

Run the unit economics with those numbers. Subscription software at a 71.8% gross margin gives about $4,100 of gross profit per location a year; the payments margin of 0.59% on $1.19M adds about $7,000 (our calculation). Toast spent about $17,500 in sales and marketing per net new location in Q2 2026 — a rough CAC proxy, not a reported metric. On software alone, that spend takes roughly 4.3 years of gross profit to recover. With payments included, it is closer to 1.6 years (our calculation).

That is the main lesson: in US restaurant software, payments subsidise the sales motion. A pure software startup selling to the same independent owner with the same field-sales model will face the same $17,500 problem without the payments margin to pay for it.

The other public vendors show the alternative path. PAR Technology serves 174,300 active sites, mostly enterprise chains, at $1,939 of annual recurring revenue per site, growing 17% (PAR, SEC filings, Q2 2026). Olo runs digital ordering for about 89,000 brand locations with dollar-based net revenue retention of 114% (Olo, SEC filings, Q2 2025). Enterprise vendors earn less per site — Toast's revenue per location is almost seven times PAR's by our calculation — but they sell once to a head office and roll out across hundreds of sites.

Where the gaps are

Hardware-style technology is already common: 71.8% of limited-service and 48.0% of full-service restaurants used self-service ordering or payment devices in 2022 (Economic Census). Software inside the business is not. Only 20.8% of food service firms used specialised software in their processes and 18% used cloud technology in 2022 (Census, Annual Business Survey). AI use was 9.4% in April 2026, with 10.9% expecting to use it within six months, and 60.5% of firms said in September 2026 they did not expect to use AI in the next six months (Census, Business Trends and Outlook Survey).

Three openings follow from the numbers:

1. Labour, not orders. The largest cost and the highest churn are in staff. Scheduling, hiring, training and retention tools address a cost that is a third of revenue, and can be priced against hours saved.
2. Embedded finance. Only 10.9% of food service firms needed financing and applied in 2022; 12% of those seeking financing went to a fintech or online lender, while 56.2% applied to a bank, and 21.5% of applicants received none of what they asked for (Census, Annual Business Survey). Lending, cash-flow tools and payment-linked advances are where a vendor earns per transaction instead of per seat — the same logic that funds Toast.
3. Direct ordering. In a 2024 consumer survey, 84% of off-premises customers said they would order through the restaurant's own website, against 71% likely to use a third-party delivery app (National Restaurant Association, proxy). The demand for a restaurant's own channel exists; the question is whether you can make it cheaper than the commission it replaces.

What the data does not tell you

• Independent margins are survey numbers. The 2.8% pre-tax figure comes from an operator survey, not tax data. Treat it as an indication.
• Franchise structure is old. The latest franchise shares are from the 2017 Economic Census. The 2022 table may change the limited-service picture.
• Labour turnover covers hotels too. JOLTS publishes accommodation and food services together.
• Vendor numbers are one company's book. Toast's $13,300 per location is the average across its own customer base; it is not the market price of restaurant software.
• Survival. For accommodation and food services together, 85.3% of new establishments from the 2024 cohort survived their first year, 59.3% of the 2020 cohort reached five years and 41.6% of the 2015 cohort reached ten (BLS, Business Employment Dynamics). Your customer base will lose a share of its locations every year regardless of your product — build that into churn.

The full table of numbers, with sources and periods, is on the Atlas page US Restaurant Market 2026 and the cell page Global restaurant tech.