Quick answer
A ghost kitchen (or dark kitchen) is a kitchen that produces only for delivery: no dining room, no waiters, no walk-in customers. It costs less than a traditional restaurant to launch (often 15,000-60,000) because you cut the fit-out, the square metres and the service staff. But the margin is fragile: delivery commissions eat 28-35% of every order and you lose high-margin sales like table drinks. It pays off only with low food cost, steady volume and numbers tracked dish by dish.
What a ghost kitchen is (and isn't)
A ghost kitchen is a food production facility built for one channel only: home delivery. The customer never sees the premises, orders through an app and receives the food at home. Everything else — dining room, frontage, footfall location — disappears.
You'll hear many names: dark kitchen, cloud kitchen, virtual kitchen, virtual brand. In practice they mean the same thing, with nuances:
- Dark kitchen / cloud kitchen: the physical structure, a kitchen with no dining room.
- Ghost kitchen: often used as a synonym, sometimes referring to the brand that exists only on the apps.
- Virtual brand: a delivery-only label launched inside an existing kitchen, even a dine-in restaurant's.
The key point is economic, not semantic: when you remove the dining room you remove half the fixed costs but also half the margin levers. No cover charge, no wine at the table, no coffee and dessert lifting the check. It all comes down to the dish and the platform's commission.
The operating models
There is no single ghost kitchen: there are at least three ways to run one, with very different economics.
| Model | What you do | Investment | Risk | |---|---|---|---| | Standalone kitchen | Rent/fit out your own unit, delivery only | Medium-high | All on your shoulders | | Slot in a fitted hub | Rent a station in a ready-made food hub | Low | High rent, less control | | Virtual brand in existing kitchen | Launch a delivery brand inside your restaurant | Very low | Cannibalisation, kitchen chaos |
The "virtual brand" model is the most underrated: if you already have a kitchen running at half capacity during off-peak hours, you can launch a second label (a poke or burger line, say) using the same equipment and staff. The marginal investment is almost zero, and every extra order covers fixed costs you're already paying.
The fitted hub (delivery food courts) lowers the entry barrier but ties you to a rent that often includes a percentage of turnover: read the contract carefully before signing.
How much it costs to open one
This is the real appeal of the model: the upfront investment is a fraction of a dine-in restaurant's. Here is an indicative budget for a small standalone kitchen (50-70 sqm).
| Item | Estimate | Notes | |---|---|---| | Deposit and initial rent | 3,000-8,000 | No prime frontage needed: a back street works | | Refit and systems | 8,000-20,000 | Extraction, electrics, plumbing to code | | Kitchen equipment | 6,000-18,000 | Used/refurbished is fine | | Paperwork, HACCP, advisors | 1,500-4,000 | Licensing, health registration | | Packaging and opening stock | 1,500-4,000 | Boxes, cutlery, first delivery | | App launch marketing | 1,000-6,000 | Sponsored slots, photos, intro discounts | | Indicative total | 21,000-60,000 | Vs 80,000-200,000 for a dine-in restaurant |
The saving doesn't come from the kitchen (that costs the same) but from everything around the dining room: furniture, customer toilets, frontage, prime location, service staff. Cut the floor space and you cut the staff: two of the heaviest line items.
But beware: having no dining room does not reduce health obligations. You still need licensing, a food safety plan, health registration and compliant premises. It is a professional kitchen in every respect.
The P&L: where the margin goes
The critical point of ghost kitchens is that they live and die on delivery, and delivery costs money. Let's run the numbers on an average order.
Assume an average order value of 25 gross, a 30% platform commission, food cost at 30% and packaging at 4%.
Gross customer price = 25.00
Commission 30% = 7.50
Food cost (30% of 25) = 7.50
Packaging (4%) = 1.00
Contribution margin = 25.00 - 7.50 - 7.50 - 1.00 = 9.00
That leaves you roughly 9.00 per order (36%) to cover rent, kitchen staff, utilities and profit. In a dine-in restaurant the same 25 dish without commission would leave over 16 of gross margin. The difference is all in the commission and the packaging.
This means one thing only: volume matters more than ever. Without a steady flow of orders, the fixed costs (rent and staff) eat that contribution margin in days. To see where you stand, start with the delivery commission calculator to fix your real margin per order.
How many orders you need: break-even
The contribution margin per order (9.00 in the example) covers the monthly fixed costs. Say:
- Rent: 1,500
- Kitchen staff (2 people): 4,500
- Utilities and sundry fixed: 1,000
- Total fixed: 7,000/month
Break-even orders = fixed costs ÷ margin per order = 7,000 ÷ 9.00 ≈ 778 orders/month, about 26 orders a day over 30 days. Below that you lose; above it you start earning.
| Orders/day | Revenue/month | Contribution/month | Gross profit | |---|---|---|---| | 18 | 13,500 | 4,860 | −2,140 | | 26 | 19,500 | 7,020 | ~0 (break-even) | | 45 | 33,750 | 12,150 | +5,150 |
The message is clear: a ghost kitchen without volume is a cash-burning machine. Before signing a lease, estimate how many orders you can realistically do in your area. To build that calculation properly, use the restaurant break-even calculator.
Multi-brand: the most powerful lever
The real structural advantage of a ghost kitchen is that you can run multiple virtual brands from one kitchen. A kitchen making burgers can, with minimal effort, also launch a poke line, a fried-chicken brand and a healthy one, each with its own identity on the apps.
The benefits:
- More visibility: you occupy more slots in the platforms' feeds.
- Same fixed costs: rent and staff don't double.
- Fast tests: launch a brand, kill it in a week if it flops.
The risks not to underestimate:
- Stock complexity: more menus = more SKUs = more waste if uncontrolled.
- Kitchen chaos at peak: four brands ordering at once can choke the line.
- Quality dilution: better two brands done well than five mediocre ones.
The rule of thumb: add a brand only if it shares at least 60-70% of ingredients with what you already produce. Otherwise your stockroom explodes.
Common mistakes
- Underestimating commissions: a "flat" 30% rarely tells the whole story once fees stack up. Operators who don't model it run at a loss without knowing.
- Thinking costs vanish without a dining room: kitchen, extraction, HACCP and production staff stay. You cut the floor, not the kitchen.
- Opening without estimating local volume: a ghost kitchen lives on order density. In a quiet area it never takes off.
- Using dine-in prices on the app: without a markup to absorb the commission, many dishes break even at best.
- Neglecting packaging: on low-value orders, boxes and cutlery easily reach 4-6%.
- Depending on a single channel: if 100% of orders come from one platform, you're at the mercy of its rules and rates. Always push a direct channel too.
Related resources
- Restaurant break-even calculator — how many orders you need to cover fixed costs
- Delivery commission calculator — your real margin per order