Ghost kitchen vs cloud kitchen: the difference

A restaurant owner is approached twice in the same month. The first pitch offers a ghost kitchen: a delivery-only brand, run from the owner's existing kitchen in its quiet hours, sold through the delivery apps under a different name. The second offers a cloud kitchen: a unit in a shared delivery hub across town, fitted out and ready, from which the owner could serve a whole new area. Both salespeople use the two terms as if they were the same. They are describing very different businesses.

The words are used loosely across the industry, and the confusion matters, because the two models carry very different costs, risks and demands on a restaurant that already has a dining room to run. One uses spare capacity in a kitchen you already pay for. The other is, in effect, a second site without a dining room.

This guide covers the difference in the comparison of ghost kitchen vs cloud kitchen, how each works, what each costs, where each goes wrong and which one an existing restaurant should consider first. Every venue and figure in it is an example.

The terms get mixed up because the industry has never agreed on them. Ghost kitchen, cloud kitchen, dark kitchen, virtual kitchen and virtual brand are all used, sometimes for the same thing and sometimes for different things, by delivery platforms, property companies and the press.

Rather than arguing about words, it helps to separate the models by what they actually are: who owns the kitchen, whether there is a dining room and how many brands cook there. Once the model is clear, the right name matters much less.

As most restaurants meet the term, a ghost kitchen is a delivery-only operation with no dining room for its brand. Often it runs from an existing restaurant's kitchen: the restaurant cooks a second menu, under a different brand name, sold only through delivery apps.

Guests ordering from the brand may never know it comes from a restaurant down the road. The restaurant gains orders from the same kitchen, the same team and the same rent, using hours when the kitchen would otherwise be quiet.

A cloud kitchen, in the sense most commonly sold to restaurants, is a kitchen facility built for delivery: a unit in a site designed for delivery-only cooking, often one of several units in the same building, each rented to a different brand, with shared loading and collection points for riders.

Operating from a cloud kitchen means opening a new kitchen in a new place, without a dining room. It brings a new rent, a new team or a split one, new equipment and the job of running a second site.

The simplest way to tell them apart is to ask one question: does the delivery brand cook in a kitchen you already run, or in a new kitchen somewhere else? If it is your existing kitchen, the model is closer to what most people mean by a ghost kitchen, or a virtual brand. If it is a new kitchen, it is closer to a cloud kitchen.

That one question decides most of the costs, risks and work involved, whatever the salesperson calls it.

A ghost kitchen run from an existing kitchen costs relatively little to start: a new menu, packaging, a brand name and logo, photography and listings on the delivery apps. The rent, the equipment and much of the team are already paid for.

Its ongoing costs are the ingredients, the packaging, any extra hours and the delivery platforms' commission, which can take a large share of each order. The question is whether what is left after those costs adds enough to the kitchen's week.

A cloud kitchen costs far more: rent on the unit, often with service charges for the facility, fit-out or equipment, a team to cook there, utilities, packaging and the platforms' commission. It is closer to opening a second restaurant than to adding a menu.

It can make sense for a brand with strong delivery demand in an area it cannot reach from its existing kitchen. It is a large commitment for a restaurant testing whether delivery works at all.

For a restaurant with a dining room, the biggest risk of a ghost kitchen is the dining room itself. A kitchen cooking delivery orders at the same moment as a full dining room can fall behind on both, and the dining room's guests, who are paying more and sitting in front of you, may wait longer.

Plan the delivery brand around the dining room's quiet hours, or limit its orders when the room is busy. A ghost kitchen that slows the dining room's service is costing more than it earns.

The best case for a ghost kitchen is spare capacity: hours when the kitchen is staffed and equipped but quiet, such as weekday afternoons or early evenings before the dining room fills. Orders in those hours use time and equipment the venue is already paying for.

Look at when your kitchen is genuinely quiet before deciding. A kitchen that is busy from noon until close has little spare capacity, however attractive the delivery orders look.

A ghost kitchen needs its own delivery menu, designed for travel: dishes that hold their temperature and texture for the length of a delivery, use ingredients the kitchen already stocks and can be cooked quickly without disrupting the dining room's service.

Not every dining room dish travels well, and a delivery menu that tries to copy the dining room often disappoints. Our guide to a delivery menu that is not your dining room menu covers how to build one.

The cheapest ghost kitchens share ingredients with the dining room. A burger brand run from a bistro that already buys beef, bread and potatoes adds little stock; a sushi brand run from the same bistro adds a whole new list of ingredients, suppliers and skills.

Choose a delivery concept that uses what the kitchen already has and does well. Shared ingredients keep waste low and prep simple.

Both models usually depend on delivery aggregators, the apps that list the brand, take the orders and send riders. They bring reach that a single restaurant could not easily build, and they charge for it, through commission on each order, fees and sometimes marketing charges.

The commission is the cost that most often decides whether delivery makes money. Our guides to delivery aggregators: the deal, read properly and aggregator commission: the real cost of an order cover what to check.

Packaging is a real cost in both models, and it is easy to underestimate. Containers, lids, bags, napkins, cutlery and labels can add a noticeable amount to every order, and better packaging that keeps food hot and intact costs more.

Cost packaging into each delivery dish's price. A dish that looks profitable before packaging can lose money once the box, the bag and the sauce pots are counted.

A delivery brand does not need the same hours as the dining room. A ghost kitchen run from an existing kitchen often works best in the hours when the dining room is quiet, such as afternoons and early evenings, and closed when the room is full.

Set the delivery hours on the platforms to match the kitchen's spare capacity, and change them when the pattern changes. A brand open during the dining room's peak may bring orders the kitchen cannot cook well.

Some restaurants run several delivery brands from one kitchen, each aimed at a different craving: burgers, bowls, desserts. More brands mean more listings and more chances to be seen, and also more menus, more packaging, more complexity and more pressure at busy times.

Start with one brand, run it well and measure it before adding another. A kitchen juggling three delivery brands and a dining room can quickly lose control of all four.

A ghost kitchen brand is not dishonest in itself, but guests increasingly want to know where their food comes from, and some are unhappy to discover that a brand with its own name and logo is cooked in a restaurant they know.

Be straightforward if asked, and make sure the food is as good as the dining room's. A delivery brand that disappoints can quietly damage the restaurant behind it if the connection becomes known.

Both models need the same food safety care as the dining room, with extra attention to packaging and timing. Hot food must stay hot, cold food cold, allergens must be labelled and packaging must keep food safe for the journey.

Allergen information matters especially for delivery, because the guest is not in front of you to ask. Make sure each dish's allergens are clear on the listing and, where required, on the packaging.

A cloud kitchen demands everything a second site does: a manager, a team, ordering, stock, cleaning, food safety records, equipment maintenance and a budget. It also lacks what a restaurant has, such as passing trade, a dining room to build loyalty and a team that sees guests enjoy the food.

It suits a business that is already strong in delivery and wants to reach a new area. It rarely suits a restaurant whose main business is its dining room.

Delivery brands live and die by their ratings on the apps. A new brand with few ratings, or a few poor ones, can disappear from guests' view, while a well-rated brand is shown more often. Ratings depend on the food, but also on things the kitchen does not control, such as the rider's timing.

Watch the ratings closely in the first months, respond where the platform allows and fix the problems that are yours: missing items, poor packaging, dishes that do not travel. A brand's rating is part of its cost and its value.

Someone has to check and pack every delivery order: the right dishes, the sauces, the cutlery, the seal on the bag. In a kitchen busy with the dining room, this job often falls to whoever is free, and that is when items go missing.

Give packing a clear owner during delivery hours, with a simple checklist for each order. A missing side or a forgotten sauce is the commonest complaint on delivery apps, and the easiest to prevent.

Measure a ghost kitchen on what it adds after all its costs: the orders, the revenue after commission, the ingredients, the packaging and any extra labour. Compare that with what the kitchen would have earned without it, which may be nothing in the quiet hours.

Watch the dining room too. If the dining room's ticket times rise, or its feedback falls, when the delivery brand is busy, the ghost kitchen's true cost includes those guests.

Measure a cloud kitchen as a separate business, with its own profit and loss: revenue after commission, against rent, team, ingredients, packaging and every other cost of the unit. Give it a budget and a date by which it must break even.

Agree in advance what you will do if it does not. A cloud kitchen that loses money every month, propped up by the restaurant, is one of the expansion mistakes that can harm both.

Before taking a cloud kitchen unit, read the exit terms. Many units come with a lease or licence of a set length, service charges and conditions about leaving the space. A delivery brand that does not work can leave the venue paying for a kitchen it no longer uses.

Prefer short terms and clear exit options, especially for a first unit. The freedom to leave if the numbers do not work is worth more than a small discount for a long commitment.

Delivery platforms decide how brands appear in their apps: where they rank, which are promoted and how quickly a new brand is seen. A new brand may start with little visibility and depend on promotions, paid for by the venue, to be noticed.

Ask the platform how new brands are shown, what promotions cost and how ratings affect placement. The cost of being seen belongs in the sum alongside the commission.

Test delivery before committing to either model. A short delivery menu from the existing kitchen, in quiet hours, under the restaurant's own name, shows whether there is demand, what the commission really costs and how the kitchen copes, at very little risk.

If the test works, a separate ghost kitchen brand may follow. If it works very well in an area the kitchen cannot reach, a cloud kitchen may be worth considering. If it does not work, little has been lost.

For a restaurant with a dining room, a ghost kitchen run from the existing kitchen is almost always the place to start, because it uses spare capacity, costs little and can be stopped easily. A cloud kitchen is a different business decision, closer to opening a second site.

The question to ask is not which term sounds more modern, but whether your kitchen has spare hours, whether delivery demand exists near you and whether the commission leaves enough behind.

Before either, look honestly at the kitchen's capacity. How many orders can it cook in an hour, at each station, without slowing the dining room? When is it genuinely quiet? Which stations are already stretched at the peak?

A kitchen whose grill is the bottleneck at dinner cannot add a burger brand at dinner without hurting the dining room. A kitchen with a quiet afternoon and a spare fryer might add a chips-and-chicken brand easily.

Delivery changes the team's work. Someone must watch the delivery tablets, pack orders, hand them to riders and deal with missing items and complaints from guests the team never sees. In a busy kitchen, these tasks fall on people who are already busy.

Decide who does what, and whether the extra orders justify extra hours. A ghost kitchen run by an exhausted team, squeezed between dining room orders, rarely lasts.

An invented bistro has a quiet kitchen between three and six on weekdays and is offered both a ghost kitchen brand and a unit in a cloud kitchen hub. The owner tests delivery first: a short menu of three dishes that share the bistro's ingredients, under the bistro's own name, only in the quiet hours, for six weeks.

Orders are steady but modest, and after commission and packaging each order adds a little. The owner launches a single burger brand from the same kitchen, limited to the quiet hours and closed when the dining room is busy, and declines the cloud kitchen. A year later, the brand adds a small, steady contribution, and the dining room's ticket times are unchanged.

servearound does not take delivery or takeaway orders and connects to no delivery platform: orders on its till and its QR menu are for seated tables, so a ghost kitchen or cloud kitchen brand's orders run through the platforms' own tablets, outside servearound. What it covers is the dining room the delivery brand must not slow down.

Kitchen analytics show ticket times day by day and by day and hour, ticket times by station, queue depth against ticket time and the items that take the most time, which show where the kitchen has quiet hours and spare stations, and where it is already stretched. Reservation analytics show covers by day and hour, so the dining room's busy times can be set against any delivery hours. See point of sale and kitchen.

As most restaurants meet the terms, a ghost kitchen runs a delivery brand from an existing kitchen, while a cloud kitchen is a separate delivery-only kitchen unit.

A ghost kitchen from your existing kitchen, because the rent, equipment and much of the team are already paid for.

It can. Run it in quiet hours, or limit orders when the room is busy.

Dishes that travel well, use the kitchen's existing ingredients and cook quickly.

It varies by platform and agreement. Read the deal and work out what each order leaves behind.

Start with one, and add more only once it works.

For a strong delivery business reaching an area its existing kitchen cannot.

Test a short delivery menu from your own kitchen in quiet hours.

Ghost kitchen vs cloud kitchen is less about words than about where the cooking happens. As most restaurants meet the terms, a ghost kitchen is a delivery-only brand run from a kitchen you already have, using spare hours and existing ingredients, while a cloud kitchen is a new delivery-only unit elsewhere, which is effectively a second site without a dining room.

For a restaurant with a dining room, start with a test from your own kitchen, a delivery menu built to travel and a clear view of what delivery aggregators leave behind after commission. Protect the dining room's service, measure what each order really adds and treat a cloud kitchen as the large business decision it is.