Skip to content
sevenweeks
Dubai · Product studio Sunday – Thursday Fixed price in AED
All posts

Building in the UAE

What it actually costs to build an MVP in Dubai

14 July 2026 8 min

Ask five Dubai studios what an MVP costs and you will get five numbers between AED 30,000 and AED 400,000. That spread is not a market failure. It is what happens when the word "MVP" is doing too much work.

Here is what the number is actually made of.

The short answer

For a first version that real customers can use — accounts, a core workflow, payments if you need them, and an admin view so you can run it — expect AED 80,000 to 150,000 in 2026, excl. 5% VAT.

Below AED 50,000 you are buying a prototype: something that demos well and cannot take a real customer. Above AED 200,000 you are usually buying either genuine complexity, or a team that has not cut the scope.

What actually drives the number

Scope, and nothing else, first

Every hour spent building is downstream of a decision. The single biggest cost driver is how many decisions are still open when the build starts.

A project with a settled scope moves at roughly twice the speed of one where screens are still being argued about in week 4. Not because anyone works harder, but because rework disappears. If you take one thing from this article: pay for the scoping, then hold the scope.

Who writes the code

A senior developer in Dubai costs roughly two to three times a junior. They also finish the same feature in a fraction of the time, and leave behind something a second person can pick up.

Agencies that quote low usually staff low. The saving is real on the invoice and imaginary in practice, because you pay it back in the second month, and again when you hire someone to take it over.

Bilingual, or not

An English-only product costs less than a bilingual one — but the gap is much smaller than people assume, if Arabic is planned from day one. Roughly 10–15% extra.

Retrofitting Arabic after launch costs far more, because right-to-left is a layout property, not a text property. Every hardcoded margin, every icon that points somewhere, every date format, every form gets revisited. See our note on what Arabic-first really requires.

Payments and compliance

Taking money in the UAE adds real work: a gateway integration, tested failure paths, and refunds. Budget AED 10,000–20,000 on top for a straightforward integration, more if you need instalments or multi-currency.

Why quotes differ so much

When you compare two proposals, check four things before you compare prices:

  1. Is the price fixed, or an estimate? An hourly estimate is not a price. It is an opening position.
  2. Who is actually building? Ask for the names, and what those people shipped last.
  3. What is explicitly excluded? A short exclusions list usually means nobody thought about it yet.
  4. Who owns the code and the accounts? If the answer is anything other than "you, from day one", the low price has a leash attached.

The costs nobody puts in the quote

These are real and they are yours, whoever builds the product:

  • Cloud and API bills. Small at first — often under AED 500 a month for an early product — but they start on day one and they are in your name.
  • Third-party licences. Payment gateway fees, email delivery, SMS, mapping. Individually small, collectively a line item.
  • Trade licence and visas, if you are also setting the company up. Independent of the software, and frequently the bigger number in year one.
  • Your own time. Seven weeks of a studio's attention needs a few hours a week of yours. Projects stall on unanswered questions far more often than on hard problems.
  • The second version. Launch is where you start learning. Keep something back for what the first month teaches you.

What you should not be paying for

  • Discovery that produces only a document. Scoping should end with decisions and screens, not a PDF.
  • Project management as a separate line. On a small senior team it is part of building.
  • A design system for a product with nine screens. You need those nine screens to be right.
  • Native apps you do not need yet. A fast web product installs from the browser and updates without review queues. Go native when something forces you to.

Should you use no-code instead?

Sometimes, yes. If you are testing whether anyone wants the thing at all, a no-code tool and a landing page will answer that in a fortnight for a few thousand dirhams. Do that first if the demand is the open question.

No-code stops paying off at a predictable point: when you need a data model that does not fit the tool, when per-record pricing overtakes the cost of building, when you need real control over performance, or when a buyer or investor asks what exactly you own. Most teams hit one of those in the first year of real usage.

The mistake is not starting with no-code. The mistake is rebuilding twice — once badly in a tool that was never going to hold it, then again properly under time pressure. Decide up front which question you are answering: does anyone want this, or can we run this as a business.

A worked example

Take a marketplace connecting two sides — the shape we see most often. A realistic first release is:

  • sign-up and accounts for both sides
  • a listing or profile each side can manage
  • search and a filtered list
  • the core transaction, with payment
  • notification email in both languages
  • an admin view so you can approve, refund, and look things up

That is roughly 14 to 18 screens once you count the empty states, the error states, and the admin. At Dubai senior rates in 2026 it lands between AED 90,000 and 130,000, built and launched.

Strip the payment and the admin view and you are near AED 60,000 — but you have also built something you cannot operate, which is usually a false economy.

What moves the number most

In order, from largest effect to smallest:

  1. Open decisions. Every unresolved question is rework waiting to happen.
  2. Number of user types. Two sides is more than twice the work of one, because every rule has to be right from both directions.
  3. Payments. Not the integration itself, but the states around it: pending, failed, refunded, disputed.
  4. Admin tooling. Easy to cut, expensive to live without.
  5. Bilingual. Real, but the smallest of the five when planned from the start.

Notice that four of the five are scope decisions, not technology decisions. That is the point: the number is mostly a function of what you decide to build, and only slightly a function of who builds it.

How to get a number you can trust

Write one page before you talk to anyone. Not a spec — a page. What it does, who it is for, what happens on the very first screen, and how you will know in month two whether it worked.

Then ask for a fixed price against that page, plus a written list of what is excluded. A studio that cannot give you one is telling you the scope is not settled yet, which is useful information on its own.

Our own starting prices are on the pricing page, in AED, with the exclusions written down. If your page does not fit seven weeks, we will say so before you pay anything.

Share WhatsApp LinkedIn
Written by Sevenweeks

Thinking about building something?

Send a short brief. We reply within one business day, Sunday to Thursday.

§ Read next
WhatsApp