Cloud Kitchen Setup in Dubai: Costs, Licensing and the Real Economics

A cloud kitchen looks like the cheap way into Dubai’s food market, and in capex terms it often is. But the format moves the risk rather than removing it: what you save on a dining room, you hand back, order by order, to the delivery apps. Whether a delivery-only kitchen makes money is decided almost entirely by the unit economics — and those are easy to get wrong when the headline setup cost is so much lower than a full dine-in opening.

This is the operator’s view of what a cloud kitchen actually is, why the capex is lower, how the licensing works, and the one reality that decides everything: aggregator commission. It is not legal or licensing advice — confirm current requirements with the relevant Dubai authority for your case — but it is the map of what matters before you commit.

The short answer

To start a cloud kitchen in Dubai you need four things in this order: a per-order margin that survives aggregator commission and packaging, a decision between a shared facility and your own unit, a trade licence plus the Dubai Municipality approvals for the kitchen, and a food-safety system run by a trained Person in Charge. What it costs is not one number but a short list of lines — licence, approvals, space, fit-out and equipment, packaging, staff — and the two that vary most are the space and the fit-out. The sections below take each in turn; the Cloud Kitchen ROI Calculator runs the margin test on your own figures in about two minutes.

What a cloud kitchen actually is

A cloud kitchen — also called a delivery-only, ghost or virtual kitchen — is a production kitchen with no dining room and, often, no public-facing storefront at all. Orders come in through delivery aggregators and your own channels; food goes out by rider. There is no host, no waiting staff, no seating area to fit out and heat and clean.

That single structural difference is the whole reason the format exists, and the whole reason its economics are different. You are running a kitchen and a brand, not a restaurant. The customer never sees the space, which is liberating and dangerous in equal measure: liberating because you can locate cheaply and build lean, dangerous because the only thing standing between you and the customer is a third-party app that charges for the privilege.

Why the capex is lower — but only the capex

The setup cost of a cloud kitchen is genuinely lower than a comparable dine-in restaurant, for reasons that are structural rather than marginal:

• No front-of-house. No dining room, no bar, no customer bathrooms, no expensive frontage. Front-of-house fit-out is one of the largest line items in a dine-in build, and it disappears entirely.
• A smaller footprint. You pay for production space, not for seating. That means a smaller unit, often in a cheaper location, because footfall is irrelevant when no customer ever walks in.
• Leaner staffing at the start. No service team to hire and train before opening — the launch headcount is the kitchen.

The honest framing is qualitative: lower, not “cheaper by a percentage.” Anyone quoting you a precise saving is guessing, because it depends on your concept, your equipment and the deal you negotiate. What matters is the shape of the cost — and the shape is unambiguously lighter on capital up front.

But — and this is the part that catches people — lower capex does not mean lower risk. It means the risk has moved from the build to the operation. A dine-in restaurant’s danger is the fit-out and the lease; a cloud kitchen’s danger is the margin on every single order. You spend less to open, then defend a thinner margin every day you trade.

What it costs: the lines to price

We do not print an AED total, because an honest one does not exist: the same brand costs very different amounts in a shared hub and in its own unit. What does exist is the list of lines every budget needs, who sets each one, and where the real figure comes from. Put a quote or an authority figure against every row before you commit.

Two lines catch first-time operators. Packaging is paid on every single order and rarely appears in the first spreadsheet. And the aggregator rate in the contract is the one that applies — model it, and model what happens if it moves.

The licensing route

A delivery-only kitchen is a regulated food business, and the approvals are not optional shortcuts. At a high level you will deal with:

• The trade licence — through the Department of Economy and Tourism (mainland) or the relevant free-zone authority, depending on your structure and how you intend to trade. The Dubai Municipality Food Code requires every food establishment to hold a valid trade licence (§2.1), which is why it heads the sequence.
• Dubai Municipality approvals for the kitchen — the Food Safety Department lists a Request for Layout Assessment for Food Establishment Licensing (service 2976) and a Request for Permit Related to Food Activities (service 2975). The Food Code asks for plans to be approved before construction (§2.2), so the assessment shapes the kitchen before a wall is built. A production kitchen is assessed just as a restaurant kitchen is.
• Food safety, a Person in Charge and HACCP — the Food Code requires a trained Person in Charge (§3.1.1), and a documented food-safety system is part of operating legitimately, not an extra. (We cover the HACCP process and timeline and Foodwatch registration separately.)

The practical question that shapes the licence is the model itself. Some operators take their own unit and licence; others start inside a shared or managed cloud-kitchen facility that provides the space and some of the approvals as part of the package. Each route has different cost, control and speed implications — and the requirements change, so verify the current ones with the relevant authority rather than relying on what was true last year. The wider sequence for any food business is in restaurant business setup in Dubai.

Shared facility or your own unit

This is the decision that most changes both the cost and the licence work, so make it deliberately rather than by default.

A shared facility suits a concept you are still testing; your own unit suits a concept whose order volume you have already seen. Our answer on whether to take your own unit or start in a shared facility goes further.

The make-or-break reality: aggregator commission

Here is the line that decides whether a cloud kitchen is a business or a treadmill. Because a delivery-only kitchen sells overwhelmingly through third-party apps, the aggregator commission comes off the top of every order — and it is large enough to turn an apparently healthy dish into a loss.

Work an order through honestly and the picture is clear. Start with the menu price. Take out the commission the app charges. Take out your food cost. Take out packaging — real money on every single order, and easy to forget. Take out your share of the kitchen’s fixed costs. What is left is the actual margin, and it is a great deal thinner than the gross margin a first-time operator assumes.

This is why the unit economics, not the capex, are the whole game:

• A dish that is comfortably profitable on the table can lose money through the app once commission and packaging are applied.
• A small change in average order value, or in the commission rate, swings the whole model — because it lands on every order, every day.
• Volume does not save a negative per-order margin; it multiplies the loss. Scaling a kitchen that loses money per order simply loses money faster.

The discipline is to know your true margin by channel and after commission, price or engineer the delivery menu to protect it, and decide deliberately how much volume you want through a third party versus your own ordering channels — where you keep more of the order but have to drive the demand yourself. We go deeper on this in delivery aggregator economics.

When a cloud kitchen makes sense

Used for the right reasons, the format is one of the sharpest tools in F&B. It makes sense when:

• You are testing a concept. Lower capex and a smaller commitment let you prove demand and unit economics before a full dine-in build. Validate first, scale second.
• You are running multiple brands from one kitchen. One production space can host several delivery brands at once, spreading fixed cost across more revenue — provided each brand’s per-order margin stands on its own.
• Your concept is genuinely delivery-suited. Food that travels well, holds its quality in a box, and sells at an order value that survives commission. Not every concept does.
• You want to enter a catchment cheaply to read real demand before committing to a flagship dine-in site there.

It makes less sense when the concept depends on atmosphere, when the average order value is too low to absorb commission, or when the food simply does not survive the journey. The format is a financial structure, not a magic trick — it rewards concepts that fit it and punishes concepts that do not.

How to start a cloud kitchen in Dubai: the order of steps

1. Model the per-order margin first. Menu price, minus commission, food cost and packaging, minus your share of fixed cost. If it is thin on paper it will be thinner in practice. The Cloud Kitchen ROI Calculator does this on your own figures.
2. Choose the model — shared facility or your own unit — from the table above, on the volume you have evidence for.
3. Start the trade licence with DET (mainland) or the free-zone authority, and confirm with that authority which steps must be complete before the licence is issued.
4. Secure the space. In a shared facility, get the written list of what the fee covers and which approvals the facility holds. For your own unit, register the tenancy contract (Ejari).
5. Have the layout assessed before you build (Dubai Municipality, service 2976), then fit out to the approved layout.
6. Apply for the food-activity permit (service 2975), put the food-safety system in place and train your Person in Charge.
7. Contract the delivery apps with the commission in writing, and cost packaging into every dish.
8. Open small and read real orders for a few weeks before adding a second brand or a second kitchen.

The operator’s checklist before committing

Before you sign for a cloud kitchen, you should be able to answer all of these on real numbers:

• What is the per-order margin after commission, food cost and packaging? If you cannot state it, you are not ready.
• What average order value does the model need to clear its fixed costs, and is that realistic for the concept and catchment?
• How many orders a day to break even? The cloud-kitchen equivalent of covers per day — and just as non-negotiable before committing.
• What is the channel mix between aggregators and your own ordering, and what does the margin look like on each?
• Is the licensing route confirmed with the relevant authority for your structure, and is the food-safety system planned in, not bolted on?
• If it works, can it scale — more brands from this kitchen, or more kitchens — without the per-order margin collapsing?

Answer those honestly and the cloud-kitchen decision is a measured one. Skip them, lured by the low setup cost, and you have bought a cheap entry into a business whose economics you never modelled.

What a disciplined cloud-kitchen launch looks like

Lower capex than dine-in, yes — but the saving is on the build, not the risk. A licence and food-safety system in place from the start. A per-order margin that survives commission and packaging. A channel mix chosen deliberately. And a concept that genuinely suits delivery. None of that is exotic; all of it is what separates a cloud kitchen that compounds from one that quietly bleeds on every order.

If you are weighing a delivery-first model, the Cloud Kitchen ROI Calculator is a two-minute, confidential way to pressure-test the unit economics — orders, average order value, food cost, aggregator commission and fixed costs — before you commit a dirham. And if you would rather have the model, the site and the approvals worked through with you, that is what our cloud kitchen consultancy in Dubai does.