What it actually costs to run a dark store in the UAE

Every dark-store pitch deck in the Gulf quotes the same headline number — a 15 to 20 minute delivery promise — and skips the line items underneath it. Before a single order ships, a dark store in Dubai or Sharjah has already committed to rent, fit-out, a trade licence, refrigeration and a technology stack. None of that is optional, and most of it is fixed whether the store does ten orders a day or a thousand.

A small dark-store unit in a Dubai industrial or residential-adjacent area typically runs somewhere between AED 80,000 and AED 180,000 a year depending on size and location, before fit-out. Fit-out itself — racking, cold room or chiller units, a packing bench, a generator or UPS for freezer continuity — adds a one-off cost that rarely comes in under AED 60,000 for a working dark store, more for anything carrying frozen or pharmacy-grade stock. A trade licence and the associated approvals (food, health, municipality depending on category) add a recurring annual line most first-time operators underestimate by half.

• Unit rent — AED 80k–180k/year depending on emirate and size
• Fit-out — racking, cold storage, packing bench, backup power
• Trade licence and category-specific approvals, renewed annually
• Insurance for stock, cold-chain and third-party delivery liability

This is where most cost models go wrong, because they price a storefront and then quietly add three or four separate tools to make fulfilment work — a picking app, a delivery-routing tool, a printer integration, a loyalty add-on — each with its own monthly fee and its own integration risk. A platform that ships the picker app, the delivery-agent app, delivery slots, substitutions and cash reconciliation as one system removes that stacking entirely. The real comparison is not "software cost" versus "no software cost" — it is one subscription against four, with one order pipeline instead of four systems that all need to agree with each other.

Fleet is where unit economics actually get decided. A single rider on a bike covers roughly 3 to 5 short-radius drops an hour in dense urban UAE traffic; a car-based rider covers fewer but can carry bulkier baskets. At AED 2,500–3,500 a month per rider plus fuel or EV charging, a five-rider fleet running two shifts is a real monthly cost before a single delivery fee is collected. This is exactly why an express-delivery charge with a free-delivery threshold matters commercially — it recovers fleet cost on small baskets while still giving it away on the orders worth keeping.

A trained picker working from a barcode scan-to-pick workflow moves a 25–30 item basket in under six minutes on a well-organised floor; without barcode picking and a clear substitution flow, that number can double, and doubled picking time is doubled labour cost per order, every single day. Substitutions matter here too — an out-of-stock item handled through an in-app customer-approval loop costs a few seconds; the same situation handled by a phone call from the counter costs minutes, and minutes are the only thing a dark store actually sells.

None of the above is where a dark store makes money — it is the cost floor. Margin comes from basket size, order density per delivery slot, and repeat rate. A wallet and reward-points programme that brings a customer back a second and third time costs almost nothing to run and directly compounds against a fixed fleet and fixed rent. Slot capacity planning, done properly, means a rider’s trip carries two or three orders in the same drop radius instead of one. Get the fixed costs right and the platform underneath it invisible, and the entire game becomes density and repeat — which is a much better problem to have than fixing four disconnected tools at 11pm on a Thursday.

Put a shape on it: a small unit running AED 130,000 a year in rent, a five-rider fleet at roughly AED 175,000 a year loaded with fuel, and one platform subscription instead of four tools is already north of AED 350,000 a year before a single picker is paid. At an average basket of AED 90 and a realistic gross margin on grocery of 18–22%, that fixed floor needs somewhere between 90 and 110 orders a day, every day, just to break even — which is exactly why density per slot and repeat rate matter more than headline order count. A dark store doing 150 scattered one-off orders a day can lose money next to one doing 100 dense, repeat-heavy orders on the same fleet.