SaaS Gross Margin Benchmarks 2025: What 172 Companies Reveal About Unit Economics

Gross margin is upstream of everything. You cannot sustain a strong Rule of 40 without the foundation of high gross margins — the math simply doesn't work. You cannot fund an effective customer success organization to drive NRR expansion if the cost of delivering your product already consumes the revenue it generates. And you cannot defend your valuation multiple through a down market if your gross margins leave no room for profitability. Every SaaS metric conversation eventually traces back to the same question: what percentage of revenue is left after you've paid to deliver the product?

The data below covers 172+ public SaaS companies as of Q1 2025, sourced from saasdb.app. The numbers show a clear distribution — and the outliers on both ends are instructive.

The Formula

Gross Margin = (Revenue − Cost of Goods Sold) / Revenue × 100

For SaaS companies, COGS includes hosting and infrastructure costs, third-party software licenses required to deliver the product, customer support costs directly tied to delivery, and professional services costs when those services are part of a bundled offering. What COGS does not include: research and development, sales and marketing, or general and administrative costs.

The most common mistake founders make when calculating SaaS gross margin is including R&D expenses in COGS. This is incorrect under standard accounting treatment and will significantly understate your gross margin. R&D is an operating expense, not a cost of delivering your existing product. Similarly, sales commissions belong in operating expenses — not in the cost of goods sold — regardless of how a deal was structured.

Gross Margin Across Public SaaS Companies

Median and Distribution

The overall median gross margin across the dataset is 77%. Best-in-class pure-software companies approach 87–88%: Adobe leads at 88%, GitLab at 87%, and HubSpot at 85%. These companies have minimal per-customer infrastructure costs, efficient delivery, and high-margin software that scales without proportional cost increases.

Infrastructure-heavy products compress significantly. Snowflake operates at 75% gross margin, and MongoDB at 74%. The reason is direct: Snowflake's cloud infrastructure costs are a meaningful component of its COGS because it manages compute and storage on behalf of customers. Every query a customer runs has an infrastructure cost that Snowflake bears. This is the trade-off of the cloud data warehouse model — massive scale and usage-based revenue, but compressed gross margins compared to pure SaaS platforms.

Domo at 67% sits at the bottom of the dataset. This reflects a combination of hosting costs and professional services revenue mixed into the top line, both of which carry lower margins than pure software. When services revenue — implementation, onboarding, ongoing configuration — is a meaningful portion of total revenue, it drags gross margin down toward the 50–60% range that characterizes services businesses, not software businesses.

Sector Breakdown

Developer tools lead with a sector median of 87% — the highest of any category. DevOps, CI/CD, and code platforms tend to be pure-software products with minimal per-customer delivery costs. GitLab at 87% is the canonical example: its software runs on customer infrastructure, not Gitlab's own, keeping COGS low.

CRM software sits at 79% sector median. This reflects mature, high-margin products like Salesforce and Adobe at the top end, offset by some mid-market tools with heavier professional services components. Security companies sit at 78% — high margins driven by software-defined security that doesn't require hardware or significant per-customer infrastructure.

Infrastructure and data companies land at 79% and 74% respectively. The data sector compresses gross margin because cloud-delivered data products (databases, warehouses, analytics pipelines) are infrastructure-intensive by nature. The hosting, compute, and storage costs required to run these products for customers are real and material — unlike a pure SaaS application where the marginal cost of an additional user is effectively zero.

What Kills Gross Margin

Three categories of cost most often compress SaaS gross margins below expectations:

Third-party AI API costs. For AI-native SaaS products built on top of LLM APIs, the cost per inference is now a meaningful COGS item. A product that makes GPT-4o or Claude API calls per user interaction can see gross margins compressed by 10–20 percentage points compared to pure SaaS. This is the central unit economics challenge for AI SaaS built on foundation model APIs rather than proprietary models. Founders building in this category need to model LLM call cost per customer per month explicitly, not as a rounding error.

Heavy cloud infrastructure. Products that manage significant compute, storage, or data pipeline infrastructure on behalf of customers — Snowflake, MongoDB, data transformation tools — carry the burden of those costs in COGS. The mitigation is architecture: products designed so that customers run infrastructure in their own cloud accounts (BYOC models) or pay infrastructure costs directly can shift those costs off the SaaS provider's P&L.

Human-in-the-loop services blended into software revenue. When a SaaS product bundles onboarding, implementation, or ongoing configuration services with the software subscription — especially if those services are significant — the blended gross margin falls toward 50–65%. Investors generally value services revenue at a lower multiple than software revenue, and the gross margin compression is one reason why. Separating professional services as a distinct line item with its own P&L is better for financial clarity, even if it complicates the pitch.

Improving Gross Margin as You Scale

Gross margin tends to expand as a SaaS business scales — but only if you engineer for it. The three highest-leverage levers:

Infrastructure optimization. Cloud infrastructure costs follow a specific pattern: early-stage companies dramatically overpay because they're on on-demand pricing, running oversized instances, and paying retail rates for every service. At $3–5M ARR, it's worth a full infrastructure audit. Reserved instances, committed use discounts, right-sizing compute, and eliminating underutilized services can reduce cloud spend by 30–50% without any product changes. That flows directly to gross margin.

Reducing support cost per customer through better onboarding. The ratio of support headcount to customer count is one of the clearest leading indicators of gross margin trajectory. If it doesn't improve as you add customers, you don't have a scalable support model — you have a services business. Self-serve documentation, in-product onboarding, and product-led activation all reduce the per-customer support burden, improving the gross margin contribution of each dollar of revenue.

Pricing structure that passes infrastructure costs through. Usage-based pricing, when done correctly, aligns customer pricing with the company's actual cost structure. If each customer's infrastructure cost scales with their usage, and usage-based pricing scales revenue with usage, the gross margin on incremental usage should be high — because the incremental cost of serving that usage is already priced in. Snowflake's model is the clearest public example: usage-based pricing with direct pass-through of compute costs preserves margin at scale despite the infrastructure-intensive nature of the product.

What This Means for Early-Stage Founders

Instrument your gross margin from your first paying customer. Most early-stage founders track revenue as a single number — monthly recurring revenue, or total revenue. This is insufficient for understanding business health because not all revenue is equal.

A company at 60% gross margin and $100K MRR has $40K of gross profit available to fund its operations. A company at 85% gross margin and $100K MRR has $85K. These are fundamentally different businesses at the same revenue level — and the difference compounds at every stage of growth.

The threshold to watch: below 60% gross margin you likely have a services business masquerading as a SaaS business. The valuation multiple you'll be assigned, the operating model you'll run, and the capital efficiency you can achieve are all different on that side of the line. If your gross margin is in the 60–70% range, identify what's driving the compression and whether it's structural (the product requires expensive delivery) or fixable (oversized infrastructure, inefficient support model, bundled services that can be repriced or removed).

The companies in this dataset that sustain 80%+ gross margins — Adobe, GitLab, HubSpot — all have the same fundamental characteristic: software that scales without proportional delivery cost increases. That is the definition of a high-quality SaaS business, and gross margin is how you measure it.

Related Benchmarks

Gross margin is one of three metrics that form the foundation of SaaS efficiency. The other two: NRR Benchmarks 2025 — what 172 public companies reveal about retention and the Rule of 40 in 2025 — how the same dataset scores on growth efficiency. If you're building an AI SaaS product and want to understand how these metrics apply to early-stage architecture decisions, Araho's AI SaaS development service is a starting point.