Business

Toast: a payments business wearing a SaaS jacket

Toast sells software to restaurants almost at cost, then makes its money on the cards their guests swipe. Read the income statement literally and you see a $6.2 billion, 26%-gross-margin company [1]. Read it the way management runs it — subscription plus the spread Toast keeps on payment volume — and you see a ~$1.8 billion recurring-gross-profit engine compounding in the high-20s%, that has just turned the corner from a decade of losses into real cash generation [2]. Understanding Toast is mostly about learning which of those two numbers to anchor on, and why.

The two numbers, side by side

Sources: FY2025 Annual Report (Form 10-K) — revenue and gross profit [3]; Adjusted EBITDA $633M [4]; free cash flow $608M [5]; GPV and Locations [6].

Source: FY2025 Annual Report, Results of Operations — Revenue, and consolidated gross profit [7].

The grey bar is what most screens see; the cyan bar is the business. The gap is payment processing recognized gross: every transaction fee Toast collects is booked as revenue, even though the bulk flows straight through to card networks and issuing banks as interchange. That accounting choice is honest [8], but it makes the company look like a low-margin reseller when it is really a software-and-spread business. Anchor on the cyan line.

How the money is actually made

Toast runs three revenue streams that could not be more different in quality. The chart below splits FY2025 by revenue and by gross profit — the second view is the one that matters.

Source: revenue by stream from FY2025 10-K Results of Operations [9]; gross profit derived from FY2025 10-K stream revenue and cost of revenue (subscription cost $264M; fintech cost $3,891M) [10].

Subscription is the prize: software access billed per location, generally on 12–36 month terms, with the rate scaling by product count and employee headcount [11]. At $936M of revenue and ~$672M of gross profit it carries a ~72% gross margin — and management says SaaS gross margin crossed 80% for the first time in early 2026 as the base scales and AI takes cost out of support [12]. Only 15% of revenue, but it is the highest-quality 15%.

Fintech is the volume engine: transaction fees charged as a percentage of each payment plus a per-transaction fee, recognized gross, and also including the fees Toast earns marketing working-capital loans through Toast Capital [13]. $5.0B of revenue converts to only ~$1.15B of gross profit — a ~23% gross margin — because most of that "revenue" is interchange passed to the networks. What Toast keeps is the spread, and that spread is the second half of the engine.

Hardware and services is, deliberately, a loss leader: $180M of revenue against roughly negative $220M of gross profit. Toast subsidizes terminals, handhelds and installation to win the location, then earns it back over years of subscription and payments. The hardware "loss" is really customer-acquisition cost in disguise — which is why payback periods, not hardware margins, are the right lens.

The real economic engine: take rate × volume × locations

Strip the accounting away and Toast's profit is a single multiplication:

Recurring gross profit ≈ (live Locations) × (payment volume per location) × (monetization take rate)

Each term has its own driver. Locations grew ~22% in 2025 to 164,000 and 171,000 by Q1 2026 [14][15]. Volume per location moves with the restaurant economy. And the take rate — how many cents of gross profit Toast keeps per dollar of guest spend — is the lever management pulls with new products and pricing.

Source: derived from reported recurring (subscription + fintech) gross profit and GPV, FY2023–FY2025 10-K [16][17]; Q1 FY2026 figure of 103 bps as reported [18].

This climbing line is the most important slide in the Toast story. The company crossed 1% of payment volume in monetization for the first time in Q1 2026 — 103 basis points, up 5 bps year over year [19]. Underneath that: a payments take rate of ~51 bps (rising as Toast optimizes processing cost and adds targeted pricing), a fintech net take rate of 61 bps including ~10 bps from Toast Capital, and the rest from SaaS [20]. Because volume itself compounds ~20%+, a rising take rate is leverage on top of leverage — the engine that turns a low headline margin into fast recurring-profit growth. The flip side: the gross processing rate Toast charges (fintech revenue ÷ volume) is only ~2.6%, so it keeps roughly a quarter of what it bills — and that spread is the single most contestable variable in the model.

The KPI scoreboard — a decade of compounding

Source: FY2025 Annual Report, Key Business Metrics (GPV $195.1B, ARR $2,047M) [21]; location counts and prior years as reported in company filings.

Three KPIs run the model and they all point the same way: locations 40k → 164k, payment volume $25B → $195B, and ARR $326M → $2.0B in five years. ARR is the cleanest forward read — it captures subscription billings plus the payments-services run-rate, and at $2.15B by Q1 2026 it is effectively the contracted base from which next year's revenue grows. Net revenue retention has run consistently above 100% (117% in 2024), meaning the average existing restaurant spends more with Toast each year through product attach and rising payments — the textbook signature of a land-and-expand platform.

From cash furnace to cash machine

For its whole life as a public company Toast lost money. That changed, decisively, in 2024–2025. The chart shows the inflection: GAAP operating income went positive, net income reached $342M, and — the number that matters most — free cash flow hit $608M [22].

Source: operating and net income from FY2025 10-K consolidated results [23]; free cash flow FY2024–FY2025 from MD and A [24]; FY2021–FY2023 FCF derived from reported operating cash flow less capital expenditures.

Two things make this real rather than cosmetic. First, stock-based compensation is shrinking as a share of the business — SBC plus payroll tax was $255M in 2025, and SBC fell to 11% of recurring gross profit in Q1 2026, roughly half its level two years earlier [25]. That matters because dilution, not cash burn, was the historical cost of owning Toast: shares outstanding roughly doubled from ~290M (2021) to ~607M (2025). Second, Adjusted EBITDA grew from $373M to $633M, and management has guided FY2026 to $790–810M with a long-term target of a 40%+ Adjusted EBITDA margin on recurring gross profit [26][27].

The balance sheet now funds the strategy: ~$2.0B of cash and marketable securities, no drawn debt, and an undrawn credit facility [28]. Capital allocation has turned to buybacks: a $250M authorization in 2024, expanded by $500M in February 2026, with ~$400M of stock repurchased year-to-date and ~$200M of authorization left as of Q1 [29][30]. A company that diluted its way through the 2010s is now retiring shares — a genuine shift.

The moat: real, but mechanism-specific

Toast incorporated in 2011 as "Opti Systems," renamed in 2012, and spent a decade building a single-vertical operating system before its 2021 IPO [31]. That focus is the moat's source. It is not a network-effects monopoly and it is not a patent wall. It is the compounding of four concrete mechanisms:

The honest read: the moat is strong in the U.S. SMB restaurant core and shallower the further Toast travels from it. In enterprise chains, international markets, and retail, Toast is the challenger, not the incumbent, and the switching-cost moat has to be rebuilt vertical by vertical. Management's own evidence is encouraging — in each new TAM, ARR is growing faster and at higher SaaS ARPU than the core did at the same age [35] — but that is a promise being kept quarter to quarter, not a settled fact.

The growth runway — four new TAMs on the same playbook

The core (U.S. independent restaurants) still grows double digits, but the bull case rests on Toast repeating its playbook in adjacent markets. Each is early; each expands the addressable location count well beyond the core.

Management frames the destination as scaling the business toward $5B and then $10B+ of recurring gross profit over the next five to ten years [40]. The right skeptical question is not whether the TAM exists — it plainly does — but whether the unit economics in drive-thru, London, and grocery ever match the beloved U.S.-independent core, where the moat is deepest.

The competitive arena — who fights for the restaurant

Toast's own 10-K refuses to name a rival, describing the field generically as "cloud-based point of sale platforms, legacy point of sale platform payments solutions, and point technology providers" [41]. The real playing field, confirmed from each rival's own filing, splits into three camps — and the comparison flatters Toast.

Sources: peer revenue / net income from each company's FY2025 filing or staged financials; Fiserv FY2025 revenue $21.2B from its 10-K [42]; market caps as of 2026-06-24 from staged peer data (Lightspeed in CAD). Toast figures from its FY2025 10-K [43]. Gaps reflect data not available in the corpus, not zeros.

Source: derived from each peer's FY2025 reported revenue and net income and staged market caps as of 2026-06-24; Block, Fiserv and Lightspeed omitted from the bubble (diversified, missing fields, or reported in CAD) and discussed in prose. Toast figures from its FY2025 10-K [44].

Read the camps, not just the rows:

• Diversified giants — Block (Square) and Fiserv (Clover) — are the most dangerous, because they can subsidize restaurant POS from vast payments franchises. But restaurants are a slice of their business, not the mission; Toast's whole-company focus on the vertical is its edge against them.
• Integrated-payments challengers — Shift4 (SkyTab) — compete hardest on processing price and are scaled and profitable, but lack Toast's software depth and SMB-restaurant density.
• Subscale restaurant-tech — PAR (Brink/Punchh), Lightspeed, NCR Voyix — are the closest pure-plays but are smaller and, in PAR's and Lightspeed's cases, still loss-making. Toast is the only large, profitable, pure-play in the category. That is the single clearest competitive fact on the page: Toast has out-executed every dedicated restaurant-tech rival on scale and now on profit.

Cyclicality, and the three things that could break the thesis

Toast is a secular grower riding a cyclical base. Its software ARR is sticky and recurring; its payments gross profit is not — it rises and falls with how much guests spend in restaurants. That exposure showed up benignly in early 2026 (GPV per location down ~1%, with management calling consumer trends "stable" and customers "resilient") [45], but a genuine restaurant recession would hit ~60% of gross profit through volume, even as the location count keeps growing.

A fourth, more technical caution: GAAP profitability is young and still partly a function of moderating stock comp. The cash is real, but the market is capitalizing only the second year of positive earnings — so multiple risk cuts both ways.

How to value it

The wrong lens is EV/Revenue: the $6B line is three-quarters payment pass-through and makes Toast look artificially cheap. The right lenses, in order:

Primary: EV / recurring gross profit (subscription + fintech gross profit, ~$1.8B in FY2025) — this is the actual revenue of the business and grew ~27% in Q1 2026 [46][47].

Secondary: P/FCF and EV/Adjusted-EBITDA — now that the business self-funds, with FY2026 Adjusted EBITDA guided to ~$800M [48].

Cross-check: take rate × GPV — model the engine directly; a few basis points of monetization on $200B+ of volume is the whole earnings swing.

Source: price and analyst targets from market data as of 2026-06-24 (company filings and consensus, as reported); market cap derived from ~607M shares at $26.30.

At ~$26 and a ~$16B market cap against ~$2B of net cash, the enterprise trades around 8× FY2025 recurring gross profit and roughly 18× FY2026 Adjusted EBITDA — a premium that demands the high-20s% recurring-profit growth and the take-rate climb continue. That is neither obviously cheap nor expensive: it is a fair price for a proven compounder where the share leadership and cash generation are now real, but where the next leg depends on take rate holding and new TAMs maturing. The stock has de-rated meaningfully from the mid-$40s a year earlier, and consensus targets ($34 mean, $35 median) sit above the current price — the market is paying for quality but has stopped paying for hope. For an intelligent investor, the work is in the take rate: if monetization keeps marching past 100 bps while volume compounds, the recurring-profit engine outgrows the multiple; if the spread compresses, the same engine runs in reverse.