Answered plainly
The cheapest company setup is a bet that nothing changes for three years
Every quotation for a low-cost UAE company prices the least representative of the three years it will live.
The quotation you were given covers the least representative of the three years
A price for a UAE company is a price for twelve months, and twelve months is the wrong unit. The licence will live longer than that, the people on it will renew on their own clocks, and the premises line will be repriced by someone who is not the person who quoted you. What decides whether a cheap setup was actually cheap is the shape of the cost across three years, and that shape is almost never drawn before a deposit is taken.
Year one is distorted twice, in opposite directions. It carries one-off charges that will never appear again — attestation of foreign documents, an external approval, the opening of an immigration file, a deposit, a fit-out — which push it up. It also carries a promotional licence price and a first-year premises rate, which push it down. Because the two distortions point opposite ways, they hide each other. A first-year total can look reassuringly close to the advertised figure while containing almost none of the recurring cost and all of the one-off cost, and the reader has no way to tell from a single number which effect dominated.
This page does not list what a package leaves out. What a low-cost UAE licence leaves out already does that, line by line, and it is the right page to read first. What follows is the other half: how those lines behave over time, which of them compound, which repeat per person, and the point at which a cheap licence stops being a cheap licence and becomes a relocation.
No figure below is a price. Licence and premises pricing in this market is promotional, revised without notice, and different for the same product in the same zone on two consecutive quarters. Anybody publishing a three-year total for a company they have not been briefed on has produced an illustration, not a quotation. Where one zone is already on your shortlist, the indicative ranges and what a headline price leaves out belong on that zone's own profile, as they do on Shams.
What each of the three years actually contains
Read the columns rather than the rows. Year one is the only column with an entry in almost every line. Year two is the first honest column, because it contains the recurring cost with none of the one-off cost and none of the discount. Year three is the column where cycles that started at different moments begin to land in the same quarter.
| Year | What lands in it | What it can tell you |
|---|---|---|
| Year one | One-off charges that never repeat, plus a promotional licence price and premises rate | Almost nothing — it is distorted twice, in opposite directions |
| Year two | The recurring cost, with none of the one-off cost and none of the discount | The honest run rate for the company you actually bought |
| Year three | Cycles that started at different moments landing in the same quarter | Whether a twelve-month average was ever safe to budget on |
Two founders, the same licence, and a different three-year shape
The single licence price is the same for both of the people below. Nothing else is.
The reason the industry quotes a licence price rather than a company price is that the licence is the only line it can know before asking who is coming. That is not sharp practice. It is the limit of what a seller controls. But it means the advertised figure describes the line that varies least, and the reader is left to model the ones that vary most.
The premises line escalates differently on each route, and this is the largest single variable
Most founders assume the premises charge behaves like rent. A tenancy does, and in Dubai its increase is tied to a published index. A desk sold with a free zone licence behaves more like a subscription, repriced from the zone's own schedule.
The asymmetry is worth stating plainly, because it cuts against the assumption that the cheap route carries less exposure. A desk sold with a free zone licence is priced from the zone's own schedule and can be revised at any renewal. A mainland tenancy in Dubai is the reverse: the increase is constrained by a published index, and the one-off costs at the front are entirely on you. A dedicated office in a Dubai free zone deserves the same reading, because the emirate's rent-increase decree names free-zone landlords too. Neither is better. They fail in different years, and the three-year model is where the difference becomes visible.
the largest single variable in the three-year model
| Judged on | A desk sold with a free zone licence | A mainland tenancy in Dubai |
|---|---|---|
| Increase constrained by a published index | No | Yes |
| Repriced at any renewal from the provider's own schedule | Yes | No |
| One-off costs at the front fall entirely on you | No | Yes |
| Behaves like a subscription rather than like rent | Yes | No |
| Reliably the cheaper route across three years | No | No |
The line that connects premises to everything else is the visa allocation. A zone may sell it as a pay-as-you-go add-on; where a zone ties it to the premises tier instead, it is not a separate purchase you can top up. Which of the two models a zone uses is on its own profile — at IFZA, for instance, the allocation follows the package rather than the floor space. Hire past the tier and the premises product has to move, which reprices the largest recurring line in the model at the moment you have least room to negotiate, because the people are already hired.
Visa cycles are per person, and they are on a different clock from the licence
A residence permit is assembled from separate transactions with separate authorities, and it is renewed the same way. At renewal the entry permit falls away, but the medical fitness screening and the Emirates ID are charged again, per person, and health cover renews on its own annual cycle regardless of the permit's term. A work permit renews with the labour file. A spouse and each child repeat the Emirates ID individually, and those old enough for the medical test, which applies from eighteen, repeat the screening too.
Two structural facts follow, and neither appears in an annual budget.
There is no such thing as an annual cost. The licence has a date. The establishment card has a date. Each permit has its own date, set by when that person actually landed. The audit has a financial-year date. The corporate tax return has a date nine months after the period ends. These cycles were started at different moments and do not converge, so cost arrives as a sawtooth rather than a line — and the peaks get taller as headcount grows. The renewal date everything else hangs on explains which of those dates actually controls the others.
Year three is usually where cycles coincide. A licence renewed annually, permits on multi-year terms, and cover renewed yearly will, at some point in the third year, fall due together. Founders who budget on a twelve-month average meet that quarter unprepared, and the common response — deferring the premises upgrade, or renewing late — costs more than the thing being deferred.
The obligations that appear on the calendar without being bought
Three of them attach to the company rather than to its trading, and a company that has issued no invoice still carries all three.
Corporate tax enrolment is not revenue-dependent. A taxable person must enrol with the Federal Tax Authority within the timeframe applicable to it, and the administrative penalty for missing that deadline is AED 10,000 — payable by a company that may have no income at all. The return then falls due whether or not the company traded.
Small Business Relief reduces the tax and not the filing. A resident taxable person with revenue at or below AED 3,000,000 in the relevant and all previous tax periods may elect it. Electing it still means enrolling, still means filing, and still means keeping books good enough to prove the revenue figure. The relief carries a sunset that has already been extended once, which is exactly the kind of date to check at the Ministry of Finance rather than to take from any page, including this one.
VAT arrives on a threshold rather than on a decision. Mandatory enrolment applies once taxable supplies and imports exceed AED 375,000 over twelve months, with the application due within 30 days of becoming liable. A trading company on a cheap licence can cross that line in a single quarter, and nothing in the licensing process warns it. What changes in the three-year model is the bookkeeping standard behind the return, and the filing cycle the authority assigns.
taxable supplies and imports over twelve months
AED 375,000the line taxable supplies and imports cross
Where the zone requires audited accounts, the audit is an annual professional cost from the first full year. A Qualifying Free Zone Person must maintain audited financial statements irrespective of revenue, so the 0% route is not the cheaper route on compliance — it is the more expensive one, bought in exchange for a rate.
The six triggers that turn a cheap licence into a move
A cheap licence is cheapest until one of these fires. Each of them is answered by a purchase larger than the difference between the cheap licence and the right one.
Every one of those is a move, not an adjustment. A move restarts the bank file, re-does the permits, and usually leaves an unexpired prepayment behind on the licence you are leaving. Moving between free zone and mainland sets out what each route actually does to the company you already have, and why nobody can quote you a price for it.
Which turns the cheapest-setup question into a single honest question. Over a three-year horizon, which of those six triggers is plausible? If the answer is none of them — a single consultant, an activity no regulator supervises, no staff, no goods, no outside capital, revenue under the VAT threshold — then the cheap licence wins on cost and there is nothing further to weigh. If the answer is one or more, the cheap licence is not a saving. It is a deferral, priced at whatever the trigger costs on the day it fires.
three years, not twelve months
- State a horizon in years, and the headcount at the end of each onePer-person lines multiply by people, not by licences
- Price the premises tier each headcount requiresThe visa allocation attached to that tier moves with it
- Get the year-two quotation in writing, government charges itemisedThe first column with no one-off cost and no discount
- Over a three-year horizon, which of the six triggers is plausible
- None of themThe cheap licence wins on cost, and there is nothing further to weigh
- One or moreNot a saving but a deferral, priced at whatever the trigger costs on the day it fires
- Which trigger fires, and in which yearA move restarts the bank file, re-does the permits and strands a prepayment
Where this three-year comparison is the wrong exercise
If the activity is not settled, modelling cost is premature. The activity decides the approvals, the approvals decide the timeline, the timeline decides the cost. Choosing an activity to fit a package is the expensive order to do this in, and no amount of modelling repairs it.
If the company needs a bank account to function, cost is not the binding constraint. A refusal is more expensive than every line in the model combined, because it costs the months as well as the fees, and the jurisdiction chosen on price is the profile that produces refusals. Read what a bank actually tests before choosing the zone, not after.
If the honest horizon is twelve months, a three-year model over-engineers a decision you intend to reverse. A founder testing a market should price the closure rather than the third renewal — striking the company off, ending the permits, settling the final return.
If the activity is supervised, the regulator's schedule is the model. Where an approval, a capital condition or a professional indemnity requirement sits in front of the licence, the licence is a rounding error in the total and comparing licence prices is measuring the wrong thing entirely.
If someone has handed you a three-year total as one number, that number is wrong. The per-person lines cannot be known until you say who is coming, and the premises line cannot be known until the headcount curve is drawn. A range with government charges itemised separately is the most anyone can honestly produce.
If the reason for the cheap licence is that the business is genuinely somewhere else, the three-year cost is not the exposure worth modelling. A company with no operating footprint here is the profile that gets exited by its bank and questioned about substance, and a low price makes that arrive sooner rather than later.
What to model before comparing two quotations
- A horizon, stated in years, and the headcount at the end of each one.
- The premises tier each headcount requires, and the visa allocation attached to each tier.
- A per-person breakdown — entry permit, screening, Emirates ID, permit issuance, health cover, work permit where applicable — multiplied by the people, not by the licences.
- Which renewal dates coincide, and in which quarter of year three.
- The year-two quotation in writing, for the same configuration, with government charges itemised separately from the service fee.
- The audit obligation the chosen zone imposes, and whether the 0% route would add one.
- Which of the six triggers is plausible inside the horizon, and the cost of answering the most likely one.
- The closure cost, if the honest plan is to test and withdraw.
Frequently asked questions
Is the cheapest licence in year one the cheapest over three years?
Frequently not, and the first-year figure cannot tell you either way. It contains one-off charges that inflate it and a promotional rate that deflates it. Compare the year-two quotations instead, for the same configuration and the same number of people, and the ranking often reverses.
Which single line moves a three-year total the most?
People, then premises — and the two are connected wherever a zone ties the visa allocation to the premises tier. Every component of a residence permit is charged per person and repeats at renewal, so a founder arriving alone and a founder bringing a family and four staff have bought the same licence and two entirely different companies.
Does prepaying several years of licence reduce the total?
Where an authority offers a multi-year term below the sum of single years, the saving is real and the lock-in is its price: you have committed to a jurisdiction for the period in which a trigger is most likely to fire. Prepay only where the six triggers have been tested and none is plausible inside the term.
What does a cheap licence actually cost when it forces a move?
More than the difference it saved, in almost every case we see. The new licence and premises are the visible part. The invisible part is the permits re-done person by person, the bank relationship rebuilt from nothing, the unexpired prepayment written off, and the weeks in which the company cannot invoice in its new name yet.
Can the licence be moved to a cheaper zone at renewal instead?
It can, and the same arithmetic applies in reverse. Moving is a new licence, new permits and a new bank conversation, so a cheaper renewal has to beat those costs rather than beat the old renewal. Where the only motive is price, the move rarely pays back inside three years. If the cheaper zone is in another emirate, what a cheap licence in the northern emirates actually costs sets out which costs travel with you.
Begin with a clear conversation
Tell us what you intend to build.
A consultant will map the structure, jurisdiction, approvals and likely cost against what you actually described — not against a package.
