SaaS Gross Margin Benchmarks: Profit Margin by ARR Band and What Buyers Expect

Short answer: the median B2B SaaS software gross margin is 80%, and 76% on total revenue including services. Top-quartile companies run 86% or better, the bottom quartile sits near 50%, and small companies under $5M ARR median 72% because fixed infrastructure is spread over less revenue. Those figures come from the 2026 SaaS and AI Performance Benchmarks report by Aleph and Benchmarkit, published 1 June 2026, covering 342 B2B SaaS and AI-native companies on full-year 2025 actuals. Last updated August 2026. Educational only, not financial advice.

What is a good gross margin for SaaS?

Anything at or above 80% on software revenue is normal, and 75% is the line most buyers and investors treat as the floor for a healthy subscription business. The 2026 benchmark data puts the median software gross margin at 80% and the median across total revenue at 76%, the gap being services, onboarding and support work that carries real human cost. Above 86% you are in the top quartile. Below 70% a buyer will want to know why before they talk about price, because a structurally low margin caps what the business can ever throw off in cash.

What is included in SaaS COGS?

Gross margin is also the fairest way to compare two software businesses built on different delivery models, which is why it matters more than the subscription label when you are weighing SaaS against an on-premise software business. Cost of goods sold for SaaS covers everything it takes to deliver the product to a paying customer: hosting and cloud infrastructure, customer support and success headcount attached to delivery, third-party software and API costs passed through in the product, and now AI inference and model costs. It does not include sales, marketing, R&D or general overhead. Those sit below the gross margin line and hit operating margin instead. Where founders most often get this wrong is putting support salaries in operating expenses, which flatters gross margin and gets corrected the moment a buyer rebuilds the P&L in diligence.

Infrastructure is the line with the most room in it. Plenty of small SaaS businesses run 10 to 15 points below their potential margin purely because nobody has looked at the cloud bill in two years, and reserved capacity, rightsizing and clearing out idle environments are the fastest way to move the number before a sale. If you have never itemized it properly, working out what your cloud and SaaS spend actually costs per customer is usually the first place a few margin points are hiding.

Does company size change SaaS gross margin?

Yes, and more than most founders expect. The median software gross margin is 72% for companies under $5M ARR and 86% for companies between $50M and $100M ARR, a 14 point spread driven almost entirely by fixed costs being spread across more revenue. A minimum viable production setup costs roughly the same at $1M ARR as at $10M ARR, and support tooling, monitoring and compliance vendors all have floors. This is one of the quieter reasons larger SaaS businesses command higher SaaS valuation multiples: they are not just bigger, they convert more of each dollar into gross profit.

What is a good net profit margin for SaaS?

For the small, bootstrapped, owner-operated SaaS that trades on marketplaces, 30% to 50% net is common and anything above 50% is strong. Venture-backed companies are a different animal and frequently run negative net margins on purpose, reinvesting gross profit into growth, which is why buyers of those businesses price on revenue rather than earnings. There is no single benchmark here because net margin is a strategy choice in a way gross margin is not. Gross margin tells you what the business is capable of. Net margin tells you what the owner decided to do with it.

Do AI costs hurt SaaS gross margin?

Not at the median, at least not yet. The 2026 benchmark data found the software gross margin median held at 80% for full-year 2025, with the authors noting that AI costs have not dented it at the median so far. The pressure shows up unevenly instead. Products where every user action triggers an expensive model call, and products priced on usage rather than seats, carry visibly thinner margins: the usage-only pricing cohort medians 62% against the 80% overall. If you are buying an AI SaaS, inference cost per active user is the number to pull apart, because it is the one cost line that grows in lockstep with the thing you are hoping to grow.

How does gross margin affect a SaaS valuation?

Two ways, and they compound. Directly, gross margin sets the ceiling on profit, so at any given revenue a higher-margin business produces more cash and supports a higher price. Indirectly, it signals what kind of business it is. An 85% margin says software. A 55% margin says there are people in the delivery loop, which means growth needs hiring, and buyers discount that. The mechanism matters if you are converting between the units different buyers use, because turning a profit multiple into a revenue multiple runs straight through your net margin: 30x monthly profit is 2.5x annual profit, which is about 1.5x annual revenue at a 60% net margin but only 0.75x at 30%. Same multiple, half the price.

How do buyers verify margin in diligence?

They rebuild it. Expect a buyer to take twelve months of raw financials and recategorize every cost line themselves rather than accept your grouping, because the definition of COGS is exactly where optimistic accounting hides. Specific things they check: whether support salaries are in COGS, whether the founder pays themselves a market wage, whether hosting is on annual commitments that expire soon, whether any third-party API contract is about to reprice, and whether free-tier and trial infrastructure has been separated from paying-customer infrastructure. Clean, consistent categorization across all twelve months is worth more than a flattering number, because inconsistency reads as either sloppiness or concealment, and both get priced in.

How to improve SaaS gross margin before a sale

Start with infrastructure, because it is the fastest and least disruptive. Rightsize instances, commit to reserved capacity where usage is predictable, delete idle environments, and move cold data to cheaper storage. Then look at support load: the goal is not to cut support but to reduce the tickets that cause it, since documentation and onboarding fixes lower COGS permanently while headcount cuts show up as a service quality problem during diligence. Audit third-party pass-through costs next, and for AI features check whether a smaller or cached model handles the majority of calls acceptably. Do this at least two quarters before you list, because buyers look at trailing twelve month figures and a margin improvement made last month barely registers in the average.

One caution: do not confuse margin improvement with cost cutting that damages the business. Buyers read a sudden drop in support spend alongside rising churn exactly the way you would expect. Margin work that survives diligence is structural, which is why retention benchmarks and margin should be reviewed together rather than one at a time.

Where margin sits in the valuation stack

Gross margin is one input among several, and it is rarely the one that decides a deal on its own. Growth rate and net revenue retention typically move a multiple further. But margin is the input that determines whether growth is worth anything, because unprofitable growth at 50% gross margin consumes cash while the same growth at 85% funds itself. If you are working out what your own business is likely to fetch, the sequence that makes sense is to establish the margin honestly first, then apply the right multiple for your size and market, then sanity check the result. Our SaaS valuation calculator handles the second step, and SDE, EBITDA and ARR multiples explains which earnings figure a given buyer will actually use. If a sale is realistic within the year, the SaaS due diligence checklist covers what else gets rebuilt alongside your margin.

Frequently asked questions

What is the average SaaS gross margin?

The median software gross margin across 342 B2B SaaS and AI-native companies was 80% for full-year 2025, and 76% measured across total revenue including services. The top quartile runs 86% or higher and the bottom quartile sits around 50%. Averages are less useful than the cut you belong to, since company size and pricing model shift the figure by more than 20 points.

Why is SaaS gross margin so high compared to other businesses?

Because the marginal cost of serving one more customer is close to zero. Once the software exists, an additional subscriber consumes a little infrastructure and some support time rather than a physical unit that must be manufactured and shipped. That is the entire economic argument for software as an asset class, and it is why buyers get nervous when a SaaS business reports margins that look like a services company.

Is 70% gross margin bad for SaaS?

Not necessarily, but it needs an explanation. At 70% you are below the 80% median and roughly at the level typical for companies under $5M ARR, where fixed infrastructure has not yet been spread thin. If the reason is scale, that improves with growth. If the reason is heavy human delivery or expensive per-call model costs baked into the product, it will not improve on its own, and buyers will price accordingly.

Does gross margin or net margin matter more when selling?

Gross margin matters more for how a buyer judges the business, and net margin matters more for the price if the business is sold on an earnings multiple. Small profitable SaaS sold on marketplaces is priced off net profit, so net margin drives the headline number directly. Gross margin is what tells the buyer whether that net figure is durable or the result of underinvestment.

What gross margin do acquirers expect from AI SaaS?

The same 75% floor most buyers apply to SaaS generally, with more scrutiny on how inference cost behaves as usage grows. Median software margin held at 80% through 2025, so AI has not moved the typical company. Where it does bite is usage-priced products, which median 62%. Expect a buyer to model your cost per active user at two and three times current volume.