How much does it cost to build an MVP in 2026?

I get this question from a founder almost every week. The honest answer is "it depends" — but that's not useful, so this post is the long form: real numbers, real ranges, the variables that move them, and the questions you should be asking yourself before talking to any builder.

For context: I'm a solo full-stack dev based in Belgium, working with clients worldwide. I've built and shipped 20+ MVPs over the last 18 months — SaaS, marketplaces, AI apps, mobile apps. I've also been on the other side of pricing as a buyer. So this is grounded in numbers I've seen quoted, paid, billed, and received.

The TL;DR pricing matrix

The headline ranges, before adjusting for any of the variables I'll cover below:

| Type | Solo dev | Freelancer / boutique | Agency | |---|---|---|---| | Landing page | €1,500–3,000 | €4,000–8,000 | €15,000–30,000 | | MVP — simple | €4,000–8,000 | €15,000–30,000 | €40,000–80,000 | | MVP — SaaS | €6,500–15,000 | €30,000–60,000 | €80,000–200,000 | | Marketplace MVP | €12,000–25,000 | €40,000–80,000 | €150,000–400,000 | | Mobile app MVP | €8,000–18,000 | €30,000–70,000 | €100,000–250,000 | | AI / RAG product | €5,000–15,000 | €25,000–50,000 | €80,000–200,000 |

These are real numbers. The agency column isn't theoretical — those are quotes I've seen founders forward to me asking "is this insane?" (Spoiler: usually yes for an MVP. Sometimes legitimate for a regulated industry build.)

Now let me explain why the ranges are so wide.

What "MVP" actually means

The first variable that moves the price 5x in either direction is what you call an MVP.

A real MVP is the smallest possible product that delivers your core value proposition to a real user, end-to-end. Not the smallest possible feature set — the smallest possible product.

That distinction matters because most founders, when they say "MVP," mean "a v1 with everything I think I need." Real MVPs strip out:

• Admin panels (do data fixes by hand for the first 100 users)
• Multi-tier pricing (charge one price, change it later)
• Onboarding tutorials (talk to your first 50 users on Loom)
• Email notifications (use a hosted tool you can set up in 20 minutes)
• Analytics dashboards (use Plausible, ship the dashboard later)
• Settings pages (hardcode for now)
• Mobile responsiveness if your users are desktop-first

A real MVP for a SaaS is: signup, the one workflow your product enables, payment for it. That's it. Three screens. Two database tables. A week of work.

A "v1" is the same product plus a dashboard, settings, billing portal, password reset, customer support inbox, and 4-tier pricing. That's three weeks. Sometimes six.

If you ask three builders to quote you an MVP and they come back at €5K, €15K, and €40K, the most likely explanation isn't that they're wildly different qualities of dev — it's that they understood "MVP" to mean three different things. Get specific on the spec before comparing prices.

The seven variables that move price

Beyond the type-of-product distinction, these are the things I weigh when scoping a project. Each of these can move the final number by 30-100%.

1. Authentication complexity

Email + password with Supabase Auth: 2 hours of work. Google + Apple OAuth: another 2 hours. SAML SSO for enterprise clients: 2-5 days. Multi-factor auth requirements: another 1-3 days. If you say "we need SSO" in the kickoff call, the price moves by €2,000-5,000.

2. Payment complexity

Stripe Checkout one-time payment: 4 hours. Stripe Subscriptions with one tier: 1 day. Multi-tier with proration, free trials, dunning, customer portal: 2-3 days. Stripe Connect (marketplaces): a full week. Multi-currency with tax handling: another week. Crypto payments: don't ask me, ask a specialist.

3. Multi-tenancy

Single-tenant (every user is independent): zero overhead. Multi-tenant with workspaces (each workspace has multiple users with roles): 2-3 days of architecture work plus careful RLS policy design. Multi-tenant with cross-workspace features (shared billing, organization hierarchies): a full week. This is where most agency budgets bloat — because it's invisible work that doesn't ship a feature, just keeps tenancy from leaking.

4. Realtime needs

No realtime: free. Notifications + live data updates (Supabase realtime channels): 1-2 days. Multi-user collaborative cursors / presence (Liveblocks or custom): 3-5 days. True realtime with sub-50ms latency for thousands of concurrent users: this isn't an MVP question, this is a senior backend engagement.

5. AI / ML integration

A basic Claude API integration with structured prompts: 1 day. RAG with vector embeddings: 2-3 days for a small corpus, 1-2 weeks for a large one with reindexing pipelines. Custom fine-tuning: rarely worth it for an MVP, but if needed, 2-4 weeks. Real-time AI agents with tool-calling and guardrails: 1-2 weeks of careful work — see my n8n + Claude post for the architecture.

6. Mobile

Web-only: no mobile cost. PWA: free, just don't pretend it's a real app. Mobile-responsive web: 1-2 days extra. React Native via Expo (iOS + Android from one codebase): see my iOS without a Mac post — adds 1-2 weeks to most builds. Native iOS + Android (separate codebases): triples the timeline.

7. Industry compliance

Consumer apps: nothing extra. B2B SaaS for SMBs: minimal. B2B SaaS for enterprise (SOC 2 readiness): adds 1-2 weeks of audit-trail and data-isolation work. Healthcare (HIPAA): another 1-2 weeks of architecture, plus auditor costs separate. Finance (PCI for stored cards, etc.): mostly handled by Stripe but adds 3-5 days of careful flow design. EU-only with GDPR: minimal extra work if you architect for it from the start, expensive if you bolt it on later.

Why solo devs are 5-10x cheaper than agencies

This isn't because solo devs are 5-10x worse. The honest reasons:

No overhead. I don't have an office, a project manager, an account executive, a director of engineering, a sales team, or a marketing team. My ratio of time billed to time productive is roughly 1.0. Agencies have a ratio closer to 0.4 because everyone working on your project also pays for everyone else.

No context-switching. I work on one project at a time, from kickoff to delivery. Agencies put 4-6 people on a project, each working 20% time. The handoffs eat 30-50% of the productive hours.

No multi-stakeholder reviews. When I have a question, I ask you directly. When an agency has a question, it goes through a project manager who emails an account executive who books a meeting with you. That meeting is 45 minutes; the question would have taken 90 seconds in Slack.

Familiar stack. I use Next.js + Supabase + Stripe for nearly every web project. I've shipped this combination 20+ times. I know exactly how long every piece takes. Agencies often pick stacks that fit the team rather than the project.

Direct ownership. When I deliver, you own the code, you have my Loom walkthrough, you can hire any other developer to maintain it. Agencies often hold the code and operate it as a managed service so they keep billing you forever.

Why agencies are sometimes worth it

Lest this become a one-sided pitch: there are real reasons to hire an agency.

If you're a regulated industry (medical devices, financial primitives, defense), the audit trail and the documentation an agency produces is genuinely valuable. If you have multiple teams that need to maintain the codebase, having an agency that documents extensively is a feature, not a bug. If your project requires specialists (a designer, a copywriter, a compliance auditor, a security reviewer), an agency that has all of them in-house can move faster than a solo dev who has to find each contractor.

If you're a Series B+ company building a critical internal tool, the agency's process maturity (kickoff documents, sprint planning, detailed acceptance criteria, formal QA) is worth the markup. Most early-stage founders don't need any of that.

The "all freelancers are unreliable" myth

Founders who've been burned by a freelancer once become convinced that all freelancers are unreliable. The truth is closer to: freelancing is a venue, not a quality signal. The 10x developers in big tech often moonlight as freelancers. So do the people who got fired from junior roles last week. The price-quality correlation is much weaker than people assume.

The reliability signal is process, not pedigree. A freelancer who shows you a live preview URL on day one, ships daily updates, and has a public portfolio of finished projects is probably going to ship yours too. A freelancer who quotes a fixed price for "two months of full-stack work" with no breakdown is probably going to ghost.

What you should actually do

If you're a founder trying to figure out what your MVP should cost, do this:

1. Write the spec yourself first. Three paragraphs. What does the product do? Who uses it? What's the one workflow it enables? Don't pay anyone to write that for you. If you can't write it, you don't know what you're building yet.

2. Get three quotes. One solo dev, one freelancer/boutique, one agency. Same spec. Note the variance.

3. Ask each one to break out the spec. "Of the things in this spec, which would you cut to ship faster?" The answer tells you whether they understand product or just velocity.

4. Pay attention to the kickoff process. A scoping call, a written summary, a discovery doc — those are signs of a builder who'll deliver. A "I'll send you a contract tomorrow" with no scoping discussion is a sign of someone who'll bill you for surprises.

5. Don't overpay for safety. A €40K agency engagement isn't 10x safer than a €6K solo dev. The variance in solo dev outcomes is wide, but so is the variance in agency outcomes. Reduce risk by writing a tight spec and gating payments to milestones, not by overpaying for the same work.

My pricing, briefly

For transparency: my own pricing for MVP builds is €6,500 fixed for a 7-day MVP (Next.js + Supabase + Stripe, ~7 hours/day of focused work). More details on the build process. For SaaS with multi-tenant complexity, I scope after a paid €1,500 discovery sprint and quote fixed-price afterward.

That number is calibrated for what I do well, in my market, at my career stage. It's not "the right price for an MVP." It's a price for this shape of MVP from this builder. Your project might need more. Find someone who can tell you with specifics why.

Three things to watch out for

Anchor-and-extract pricing. A builder quotes a low number to win the engagement, then bills you for "scope changes" every week until you're at 3x the original quote. Defense: get a fixed-price quote with a written scope document signed by both sides.

Hourly billing without a cap. Hourly is fine for ongoing maintenance. For a fixed deliverable, hourly billing transfers all the risk to you. Defense: insist on fixed-price for fixed-scope work.

The "we'll figure it out as we go" approach. No spec, no timeline, no milestones. This isn't agility — it's a setup for unbounded billing. Defense: even rough scoping is better than none.

What I'd tell my younger founder self

If I went back five years and gave myself one piece of MVP-cost advice, it would be: the cheapest MVP is the one that gets you to your first paying customer fastest, not the one with the lowest price tag.

A €3,000 MVP that ships in two weeks and finds you ten paying users is dramatically better than a €15,000 MVP that ships in three months and is technically nicer. Optimize for speed-to-revenue, not for hours-of-engineering.

If your MVP question is "what does it cost?", you're already asking the wrong question. The right question is "what's the smallest thing I can ship that proves my hypothesis?" Once you have a clear answer to that, the cost question usually answers itself.

---

Looking for an MVP build? Read about my 7-day MVP service, see the case studies, or start a project. Reply within 2 hours during EU business hours.