How much does an MVP cost in 2026? A worked estimate

The short answer

An MVP’s build cost comes down to one multiplication:

people × weeks × rate, plus the third-party services it runs on.

The scope decides the people and the weeks; where and how you hire decides the rate. That is why published “MVP costs” range so widely: they multiply different scopes by different rates and rarely say which. In the worked example below, a focused B2B web MVP takes about 25 person-weeks of design, engineering, testing and delivery work. At the three illustrative rates this guide uses, that is roughly $40,000 to $150,000 to build, before running costs: the same scope, priced by who builds it.

We don’t publish a price list ourselves, because the scope decides the number; after a short discovery, our custom software and SaaS team gives a written estimate with the team, timeline and assumptions spelled out. This guide shows you how to build that estimate yourself, so you can sanity-check anyone’s quote, including ours.

What an MVP is, and what it isn’t

Eric Ries, who popularised the term, defines a minimum viable product as “that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort.” He adds: “MVP, despite the name, is not about creating minimal products.” The goal is learning, not a small version of the final product.

Y Combinator’s Michael Seibel is blunter about scope: “you should be able to build it fast, in weeks, not months.” If the list won’t fit, “just cut the stuff that clearly isn’t important. And if there’s no non-important things, start cutting important things.” He also names the exception: in heavily regulated industries such as insurance or banking, “it’s hard to launch” something that small.

In practice, a good MVP:

• does one job, for one type of user, end to end;
• is solid enough that people trust it with real work (and real data);
• measures whether that job gets done, so the next release is a decision, not a guess.

It is not a prototype that falls over when used, and it is not the full roadmap with the edges trimmed. The difference matters for cost: a prototype can be thrown away, while an MVP is the foundation you keep building on.

What drives the cost

Six things move an MVP estimate more than anything else:

1. The number of core workflows. Each distinct job a user can do (book, approve, invoice, report) needs screens, rules, permissions and tests.
2. Platforms. A responsive web app is one codebase. Adding iOS and Android adds a second client, even with a cross-platform framework.
3. Integrations. Each external system (payments, calendars, a CRM, an ERP) adds mapping, error handling and testing. Poorly documented APIs add more.
4. Accounts and tenancy. Single users are simple; organisations with roles, invitations and separated data are not.
5. AI features. Calling a model is quick; making its output reliable on your data, with evaluation and a fallback, is real work.
6. Compliance. Health, financial or children’s data brings requirements (audit logs, encryption, access reviews) that have to be designed in.

Effort by feature: a planning table

The table below is the effort model this guide uses. The ranges are planning assumptions for a small senior team building with a mainstream web stack and managed services, not measurements or quotes; your scope will differ. Effort is in person-weeks (one person, one working week) and includes design, engineering and testing for that piece.

Turning weeks into money: rates

The same 25 person-weeks can cost very different amounts depending on who does the work. Two kinds of public figure help calibrate a rate, and it is worth knowing the difference:

• Employee pay. In the US, the median software developer earned $135,980 a year in May 2025, about $65 an hour (US Bureau of Labor Statistics). In the UK, the provisional 2026 median for programmers and software development professionals is £58,770 a year, or £31.10 an hour (Office for National Statistics). An agency or contractor rate has to cover far more than salary: benefits, payroll taxes, unbilled time, tools, management and margin, so it sits well above salary divided by hours.
• Supplier rate benchmarks. Accelerance’s 2026 survey of 60 outsourcing partners puts senior developers at $31–$41 an hour in Asia, $60–$75 in Latin America and $64–$76 in Europe. It publishes no US onshore figure, and Accelerance itself cautions that “hourly rates are a poor measure of the true cost of software development”: a cheaper hour that takes twice as long, or needs rework, isn’t cheaper.

For the worked example we use three illustrative rates: $40 an hour (the top of Accelerance’s senior band for Asia), $75 an hour (the top of its senior bands for Latin America and Europe), and $150 an hour for an onshore US team. The last is our assumption alone: we found no current public benchmark for US agency rates we would trust, so replace it with real quotes. Treat them as reference points for the arithmetic, not as anyone’s price list, including ours.

A worked example

The product (hypothetical): a web app for small service businesses to schedule jobs, assign them to staff and send invoices. Customers sign up as organisations, invite their team and pay a monthly subscription.

Discovery and product design

3

Project foundations

1.5

Sign-up, sign-in and roles

1.5

Organisations and teams

2

Three core workflows (jobs, scheduling, invoices) at 2 each

6

Subscription billing

1.5

Email notifications

1

Admin screens

1.5

One integration (calendar sync)

1.5

Subtotal

19.5

Testing, hardening and delivery (25%, rounded)

5

Total effort

24.5, call it 25 person-weeks

A team of three to four people (a product designer for the early weeks, two engineers, part-time QA and a delivery lead) gets there in roughly eight to ten calendar weeks. At 40 hours a week, 25 person-weeks is 1,000 hours.

Effort

25 person-weeks × 40 hours = 1,000 hours

At $40 an hour

$40,000

At $75 an hour

$75,000

At $150 an hour

$150,000

Build cost, before running costs

$40,000 to $150,000

Two notes on reading it. The range comes entirely from the rate: the scope is identical, which is why comparing quotes only makes sense once you know the hours behind them. And this example is a fairly complete first product, with organisations, billing and an integration. An MVP built only to test demand can be much smaller, closer to the “weeks, not months” Y Combinator recommends, and cutting scope is the cheapest saving there is.

Our AI CRM SaaS case study shows what sits further along the same road: multi-tenant workspaces, a no-code workflow builder and AI features. That is a platform, not an MVP, and the planning table explains why: each of those capabilities is several building blocks on its own.

What it costs to run after launch

Building is the larger cost, but an MVP also needs paying for every month. For a small product on managed services:

Free tiers carry most MVPs through their first users. The cost that does not go away is maintenance: dependency and security updates, bug fixes, and small improvements users ask for. Budget engineering time for it every month, not just money for hosting.

How to spend less without building something fragile

• Cut scope, not quality. Remove a workflow before you remove the tests. Fragile software costs more to fix than it saved.
• Buy the commodity parts. Authentication, payments, email and file storage are solved problems; use managed services.
• Start on one platform. A responsive web app usually comes first; see web app vs mobile app for when that’s wrong.
• Fix the price of the first phase. A short, fixed-price discovery turns open questions into a plan before the larger spend; see fixed price vs time and materials.
• Measure from day one, so the second release is spent on what users actually do.

Sources

Every source below was read on 26 October 2026. Prices and product details change; check the current page before relying on a figure.

1. Eric Ries, Startup Lessons Learned: Minimum Viable Product: a guide
2. Y Combinator: How to plan an MVP (Michael Seibel)
3. US Bureau of Labor Statistics: Occupational Outlook Handbook: Software Developers, Quality Assurance Analysts, and Testers
4. O*NET OnLine (US Department of Labor): Local wages: Software Developers (15-1252), BLS 2025 data
5. Office for National Statistics: Occupation (4-digit SOC) ASHE: Table 14, 2026 provisional
6. Accelerance: 2026 Outsourcing Rates: Global Costs Are Trending Down
7. Vercel: Vercel Pricing
8. Supabase: Pricing & Fees
9. Neon: Pricing
10. Clerk: Pricing
11. Stripe: Pricing & Fees
12. Apple: Apple Developer Program
13. Google: Get started with Play Console