How Much Does It Cost to Run a SaaS: The Real Monthly Number (2026)

Most founders brace for a scary AWS bill and then get blindsided by everything else. Infrastructure is almost never the line item that sinks a SaaS budget, payroll, tools, and payment friction are.

The Short Answer: What Running a SaaS Actually Costs

There is no single number, but there is a usable range once you separate SaaS companies by stage. The mistake most founders make is comparing their burn to a company three stages ahead of them.

A realistic monthly range by stage (pre-revenue, early, scaling)

These ranges are directional, not a quote. A two-person bootstrapped SaaS and a seed-funded team with the same product can differ by 10x in monthly burn because one is paying salaries and the other is living on savings.

Why "it depends" is a cop-out, and the variables that actually move the number

"It depends" is technically true and practically useless. The variables that actually move your number are headcount, whether you pay yourself, your pricing model (usage-based tools tend to have spikier infra costs than flat-fee SaaS), and how much you spend to acquire each customer. Get specific about those four and the range above gets a lot narrower for your situation.

The one metric that matters more than total cost: cost per active user

Total monthly spend tells you almost nothing in isolation. Cost per active user, your total burn divided by monthly active accounts, is what lets you compare your own SaaS month over month and against public benchmarks like the ones published by SaaS Capital and OpenView's SaaS Benchmarks. A SaaS spending $40,000 a month with 2,000 active users is in a very different position than one spending $40,000 with 200.

Infrastructure and Hosting: Smaller Than You Think

Founders fixate on the cloud bill because it is the most visible number on a dashboard. For most SaaS companies under $5M ARR, infrastructure is 5 to 15 percent of total spend, not the dominant cost.

Compute, storage, and bandwidth: where the dollars go

Compute (the servers running your app) is usually the largest infra line, followed by storage and bandwidth. AWS's EC2 pricing page is a reasonable proxy for compute costs across providers, since Google Cloud and Azure track closely on raw instance pricing. Storage is cheap until you are hoarding logs or user-uploaded files without a retention policy.

Serverless vs. managed vs. self-hosted cost curves

There is no universally "cheapest" option. The right call depends on whether you are optimizing for dollars or for engineering hours, and most early SaaS teams should optimize for hours.

The egress-fee trap and why cloud bills spike without warning

Data egress (the fee to move data out of a cloud provider) is the most common source of a surprise bill. It rarely shows up in back-of-envelope estimates because compute and storage calculators do not prominently feature it. AWS's own data transfer pricing page is worth reading end to end before you commit to a provider, and the widely cited a16z piece on "The Cost of Cloud, a Trillion Dollar Paradox" is a good primer on how these costs compound at scale.

Payroll: The Line Item That Dwarfs Everything Else

If infrastructure is 5 to 15 percent of spend, payroll is usually the other 60 to 80 percent. This is the number most "SaaS cost" calculators quietly ignore because it is harder to generalize than a server bill.

Why engineering salaries are 60 to 80 percent of true SaaS cost

Software is a people business wearing a technology costume. The U.S. Bureau of Labor Statistics wage data for software developers gives a sense of the baseline cost of a single senior engineer before benefits, payroll tax, and equipment are added. Multiply that by even a small team and payroll overtakes every other line item combined.

Founders' unpaid labor is a real (hidden) cost

If you are not paying yourself a market salary, your SaaS is not actually cheap to run, it is being subsidized by your own unpaid labor. This matters when you compare your burn rate to a funded competitor's, because their number includes a founder salary and yours might not. Paul Graham's essay on being "default alive" versus "default dead" is a useful mental model for why this distinction changes your runway math, not just your spreadsheet aesthetics.

Contractors vs. full-time: the break-even math

Contractors cost more per hour but carry none of the overhead of benefits, payroll tax, or severance risk. The rough rule of thumb is that a contractor needs to be busy on your project for roughly 6 months or less before a full-time hire becomes the cheaper option, assuming comparable skill and output. Below that threshold, contractors are usually the more capital-efficient choice.

The Third-Party Tool Stack Nobody Budgets For

Most SaaS teams can name their AWS bill from memory and have no idea what they spend across their tool stack until finance adds it up. This is often a $1,000 to $5,000 monthly surprise.

Analytics, error tracking, email, support, CRM: the creeping line item

None of these feel expensive individually. Added together across a growing team, they regularly hit four figures before anyone notices.

Per-seat pricing vs. usage-based pricing traps

Per-seat tools (most CRMs and support desks) scale with headcount, which means they get expensive exactly when you are growing fastest. Usage-based tools (most analytics and email platforms) scale with customer activity, which means a successful product launch can trigger a cost spike with no warning. Know which category each tool in your stack falls into before you sign an annual contract.

An audit framework to cut 20 to 30 percent of tool spend

A simple quarterly audit catches most of the waste: list every tool with a login count and last-used date, cancel anything with zero active logins in 60 days, downgrade seat-based tools to match actual headcount rather than peak headcount, and renegotiate annual contracts at renewal instead of auto-renewing. Teams that run this exercise consistently find 20 to 30 percent of tool spend was dead weight.

Payment Processing and the Margin You're Quietly Losing

Payment processing fees are one of the few costs that scale directly with revenue, which means they feel small early and become one of your largest line items at scale.

Stripe-style fees: the 2.9% + $0.30 that compounds at scale

The standard online card rate, popularized by Stripe's pricing page, is around 2.9 percent plus $0.30 per transaction for most SaaS billing. On $50,000 of monthly recurring revenue that is roughly $1,450 to $1,600 a month, every single month, forever, unless you negotiate volume pricing or switch processors.

Failed payments, dunning, and involuntary churn costs

Card declines are not just a support headache, they are a direct revenue leak. Expired cards, insufficient funds, and bank fraud flags cause a steady trickle of involuntary churn that dunning tools like Stripe's own retry logic, or dedicated services, can recover but never eliminate. Budget for this as a percentage of revenue loss, not a one-time fix.

Chargebacks, taxes, and merchant-of-record overhead

Chargebacks carry both the refunded amount and a flat penalty fee, and tools like Stripe Radar help but do not eliminate the cost. Sales tax and VAT compliance across jurisdictions is its own overhead, which is why many SaaS companies route payments through a merchant of record that absorbs tax liability in exchange for a higher processing cut.

Customer Acquisition Cost: The Cost of Getting Users in the Door

CAC is often tracked separately from "operating costs" in founder spreadsheets, which is a mistake. It is as real a monthly expense as payroll or hosting.

CAC as an operating cost, not a growth line

If you spend $10,000 a month on ads and content to acquire customers, that is $10,000 of operating cost, full stop. Treating it as a discretionary "growth" line that can be turned off without consequence ignores that most SaaS pipelines slow down within weeks of a spending freeze.

Paid vs. organic cost structures

Paid acquisition (search ads, social ads) has a cost structure that scales linearly with spend and shows results in days. Organic acquisition (content, SEO, community) has high fixed cost and a long lag before payoff, but a much lower marginal cost per customer once it compounds. Most healthy SaaS cost structures blend both rather than betting everything on one.

How CAC payback period changes your runway math

CAC payback period, how many months of revenue from a customer it takes to recover what you spent acquiring them, is the number that should actually govern your spending decisions. A SaaS with a 6-month payback can reinvest aggressively; one with an 18-month payback needs to slow down and extend runway. Benchmarks from firms like OpenView and SaaS Capital are a reasonable gut check for where your own payback period should land relative to peers.

Put the Numbers in Your Own Spreadsheet

Every range in this article is a starting point, not your answer. The only number that matters is the one built from your own infra invoices, payroll register, and tool subscriptions.

Build your own worksheet

Pull your last three months of actual spend across every category covered here: infrastructure, payroll (including unpaid founder time at market rate), tools, payment processing fees, and customer acquisition cost.

Plug in your own figures

Divide each category by your current month's revenue to see it as a percentage, not just a dollar figure. Percentages are what let you compare month over month even as your absolute revenue changes.

Benchmark your cost per user against the ranges above

Once you have a true monthly total, divide it by active users to get your real cost per active user, and compare it against the stage ranges in the first section. That single number will tell you more about the health of your SaaS than any individual line item.

Hidden and One-Time Costs Founders Forget

Beyond the recurring monthly categories, every SaaS eventually hits a set of costs that show up rarely but hit hard when they do.

Compliance, SOC 2, and security audits

Once you start selling to mid-market or enterprise customers, a SOC 2 report stops being optional. The process, covered in detail by compliance platforms like Vanta, typically involves both a tooling subscription and an auditor fee, and it recurs annually, not once.

Legal, accounting, and insurance

Incorporation, contract review, bookkeeping, tax filing, and business insurance (including, eventually, cyber liability coverage) are easy to underbudget because they are irregular rather than monthly. A reasonable annual reserve for a small SaaS is a few thousand dollars at minimum, scaling up fast as headcount and customer contracts grow.

Downtime, incident response, and technical debt interest

An outage costs more than the hour it takes to fix, it costs support time, customer trust, and sometimes contractual SLA credits. Technical debt behaves like financial debt: ignoring it does not make the cost disappear, it just adds interest in the form of slower future development.

A Worked Example: Full Monthly Cost of a Real Early-Stage SaaS

Abstractions only go so far. Here is a plausible full monthly cost breakdown for a 2-person SaaS with roughly 500 paying customers at an average of $40 per month, or $20,000 in MRR.

Line-by-line monthly P&L

At $20,000 in MRR against $29,400 in monthly cost, this business is burning roughly $9,400 a month, which is a completely normal position for an early-stage SaaS still proving its growth motion, as long as there is enough runway to cover it.

Where each dollar goes as a percentage

Payroll alone is nearly two-thirds of total spend, which matches the broader pattern across the stage ranges in the first section. Infrastructure, the line item most founders worry about first, is the smallest percentage on the entire table.

What changes at 5,000 customers

Scale this example to 5,000 customers and the shape of the budget shifts: infrastructure grows but stays a small percentage, payroll grows faster as support and engineering headcount expands, and customer acquisition cost often becomes the largest single line item as paid channels get more competitive and organic growth alone can no longer sustain the pace.

Frequently Asked Questions

How much does it cost to run a SaaS per month? It ranges from a few hundred dollars for a solo pre-revenue founder to hundreds of thousands of dollars a month for a scaling, venture-funded team. The number that matters more than the raw total is cost per active user, which lets you compare your spend against your own growth rather than against an unrelated company's budget.

What is the single biggest cost of running a SaaS? Payroll, almost always. Engineering, support, and sales salaries typically make up 60 to 80 percent of true monthly cost once you account for founders' time at a fair market rate, far outweighing infrastructure or tooling.

How much should I budget for infrastructure vs. payroll? A reasonable early-stage split is roughly 5 to 15 percent of total spend on infrastructure and 60 percent or more on payroll, with the remainder spread across tools, payment processing, and customer acquisition. The exact ratio shifts as you scale and as your product's compute intensity changes.

Can you run a profitable SaaS as a solo founder with low costs? Yes, it is one of the more common bootstrapped SaaS patterns. Keeping the tool stack lean, using serverless or managed infrastructure instead of hiring DevOps help, and relying on organic acquisition over paid ads can keep total monthly cost in the low thousands while still generating meaningful revenue.

What hidden SaaS costs catch founders off haven't budgeted for? Compliance work like SOC 2 audits, payment processing chargebacks and failed-payment churn, cloud egress fees, and the opportunity cost of unpaid founder labor are the most commonly underbudgeted items. None of them show up on a basic hosting bill, which is exactly why they surprise people.

How do I calculate my cost per active user? Add up every monthly cost category covered in this article (infrastructure, payroll, tools, payment processing, customer acquisition, and a reserve for compliance or one-time costs), then divide the total by your count of monthly active accounts. Track it over time rather than as a single snapshot, since the trend line matters more than any one month's number.

The real monthly cost of running a SaaS is rarely the number on your cloud invoice. It is payroll, tools, payment friction, and acquisition spend added together, measured against how many active users that spend actually supports. Build the worksheet with your own real numbers, and you will have a far more honest picture than any industry-wide range can give you.