How do ghost kitchen business models actually work?
In our ghost kitchen market deck, you will find everything you need to understand the market
SUMMARY
Ghost kitchen business models work when they either sell low-cost kitchen capacity to restaurant brands that already have demand, or use shared kitchens to make strong owned or licensed brands more productive. The weakest version is still the unknown delivery-only brand that depends on marketplaces for nearly every customer.
The pandemic did not kill restaurant delivery. It killed the assumption that delivery growth would automatically fill every new ghost-kitchen site. DoorDash orders are still growing, while Kitchen United, REEF and several dedicated cloud-kitchen networks have sharply reduced or abandoned physical capacity.
The biggest advantage is upfront capital, not magically high restaurant margins. Removing the dining room, front-of-house fit-out and prime retail lease can make a new location far cheaper to open, but food, labor, packaging and delivery-platform commissions still sit in the operating P&L.
Customer acquisition can matter more than kitchen rent. DoorDash and Uber Eats can provide demand, but marketplace delivery fees commonly run from roughly 15% to 30%, and an unknown virtual brand may also need promotions just to get noticed.
Multi-brand kitchens are more convincing because one rent bill, one team and one set of equipment can support several demand pools. The catch is that extra brands only help when they bring genuinely incremental orders without adding too many ingredients, processes and operational headaches.
Owning the brand is usually more valuable than owning the kitchen box. Kitopi, Rebel Foods and Wonder have all moved deeper into brand ownership, restaurant operations or direct consumer relationships because demand is harder to replace than production capacity.
Virtual-brand franchising can be highly capital-efficient because local operators fund much of the kitchen and labor while the brand owner earns royalties or revenue share. Its weak point is quality control: the faster a brand spreads across independent kitchens, the harder it becomes to guarantee that customers receive the same product everywhere.
The industry is also moving back toward visible storefronts. Taster, Kitopi, Rebel Foods and Wonder increasingly combine delivery with pickup, dine-in or food-hall formats because physical locations add customer discovery and create lower-cost transactions than marketplace delivery.
Ghost kitchens removed the need for prime pedestrian real estate, but they did not remove geography. Dense delivery zones still matter because order volume, courier travel time and kitchen utilization all deteriorate when customers are spread too far apart.
Technology helps, but it is mostly a margin tool. Routing orders, forecasting demand and standardizing kitchen work can save a few important percentage points, yet software cannot rescue a concept that customers do not deliberately choose.
The durable ghost-kitchen model is therefore restaurant infrastructure attached to real demand. Strong brands, dense local demand, shared capacity and several customer channels can produce attractive economics; an anonymous brand paying a marketplace to supply every order usually cannot.
What does a ghost kitchen actually sell?
A ghost kitchen can make money from kitchen rent, food sales, restaurant royalties or operating services, so the term covers several very different businesses.
CloudKitchens is closest to a specialized landlord. Restaurant operators rent equipped kitchen units and get shared infrastructure such as storage, utilities and courier handoff. CloudKitchens gets paid whether the tenant sells burgers, pizza or salads, although weak restaurant sales eventually become its problem if tenants leave.
Rebel Foods sits much closer to the restaurant side. It owns concepts such as Faasos, Behrouz Biryani and Oven Story, runs the kitchens and collects revenue from the food itself. Taster takes another route: it creates restaurant brands, increasingly with online creators, then lets franchisees and licensees operate them. Kitopi started by cooking for third-party restaurant brands but now owns and operates a large portfolio itself.
These models can look identical on Uber Eats. Economically, they barely resemble one another. One company wants high kitchen occupancy. Another wants food gross margin. Another wants royalties. Another wants several brands to share the same cooks and equipment.
That is the first thing to get right before asking whether ghost kitchens “work.”
| Ghost kitchen model | What it actually sells | Main way it makes money | Biggest risk |
|---|---|---|---|
| Shared kitchen operator | Equipped kitchen capacity | Rent and service fees | Empty kitchens |
| Multi-brand restaurant operator | Meals | Food sales | Food, labor and delivery costs |
| Virtual-brand franchisor | Restaurant brand and operating system | Royalties or revenue share | Weak demand or poor execution |
| Managed kitchen operator | Restaurant operations | Fees, revenue share, food sales | Operational complexity |
| Integrated food platform | Food, ordering and delivery | Several layers of the transaction | Very high capital needs |
Did ghost kitchens really collapse after the pandemic?
The pandemic-era ghost kitchen land grab collapsed, while restaurant delivery itself is currently doing just fine.
DoorDash's latest quarterly results make the split pretty clear. Total orders increased 27% from a year earlier to 970 million and marketplace gross order value reached $33.1 billion. Deliveroo's acquisition inflated those figures, but even excluding Deliveroo, DoorDash orders were still up 17% and gross order value was up 23%. DoorDash also said growth in its U.S. restaurant business accelerated slightly.
Consumers therefore did not suddenly abandon delivered food once restaurants reopened. The National Restaurant Association's latest off-premises study found that 37% of U.S. adults order restaurant delivery at least weekly, while nearly three quarters of overall restaurant traffic now happens away from the dining room through delivery, takeaway and drive-thru.
What disappeared was a lot of the infrastructure built around unrealistic expectations.
Kitchen United is the clearest example. It raised $100 million in 2022 from investors including Kroger, Restaurant Brands International and Simon Property Group, and at one point planned hundreds of locations. By late 2023 it was selling or closing every physical site and going back to software.
REEF's partnership with Wendy's went through an even more dramatic reset. Wendy's initially discussed as many as 700 delivery kitchens with REEF. The plan was later cut to roughly 100 to 150 before Wendy's closed its remaining U.S. REEF kitchens.
Singapore gives us a broader market test. The country went from 18 cloud-kitchen facilities at the pandemic peak to 11, while Grab and Deliveroo both shut their dedicated cloud-kitchen operations there.
Delivery demand survived all three cases. Too much ghost-kitchen capacity did not.
This market map, featured in our ghost kitchen market deck, highlights top companies and startups in the ghost kitchen market
Does renting ghost kitchens to restaurants still work today?
Kitchen-as-a-service still works today when it helps restaurants with proven demand open locations more cheaply and quickly.
CloudKitchens is the strongest evidence that the rental model itself survived the shakeout. The company currently advertises more than 90 U.S. facilities across 55 cities and says more than 2,000 kitchens are operating across the U.S. and Canada. Its website now prominently features established chains such as Popeyes, Five Guys, Papa Johns and Burger King.
The customer mix shows how the proposition has matured. A known restaurant already has a menu, customers and operating history. The ghost kitchen mainly solves real estate, construction and delivery coverage.
CloudKitchens says established brands inside its facilities generate median annual sales of about $1.6 million per location and can open for less than 10% of the upfront investment required for a traditional restaurant. Those are company-reported marketing figures, so they should not be treated as an independent industry benchmark. Still, they explain what CloudKitchens is selling these days: return on capital and speed of expansion.
Nimbus shows how the same idea can work at smaller scale. Food businesses can rent kitchens by the hour or take dedicated space for longer periods. A caterer, packaged-food company, bakery or delivery restaurant can therefore buy only the production capacity it needs instead of signing a conventional restaurant lease.
The failed version was closer to “build kitchens everywhere and assume thousands of new virtual restaurants will fill them.” Kitchen United showed how quickly that thesis can unravel. The more durable version starts with an existing food business that already knows where its customers are.
If you want more recent data on this point, please see our latest ghost kitchen market report.
Are ghost kitchens actually much cheaper than normal restaurants?
Ghost kitchens are usually much cheaper to open than full restaurants, but their operating-cost advantage can shrink quickly once delivery enters the P&L.
The upfront saving is real. A delivery kitchen needs no large dining room, expensive customer-facing fit-out or full front-of-house team. Rebel Foods has previously said that when it moved Faasos away from conventional restaurants, rent fell from roughly 15% of sales to around 4%. CloudKitchens currently claims an established restaurant can enter one of its locations for less than one tenth of normal restaurant opening investment.
That can transform the return on capital. Spending $150,000 to reach a new trade area creates a very different hurdle from spending $1.5 million.
The recurring expenses are harder to escape. The National Restaurant Association's latest operating-cost study, based on more than 900 U.S. restaurant operators, found that a median limited-service restaurant spent 32.4% of sales on food, 31.7% on labor and 5.2% on occupancy. Combined, that is 69.3% of revenue before utilities, insurance, technology, packaging, marketing and other expenses.
Limited-service restaurants in that study produced median pretax income of only 4% of sales.
A well-run ghost kitchen should beat the typical restaurant on occupancy and may use labor more efficiently, especially when several brands share the same site. The problem appears when those savings are replaced by a large marketplace commission and more packaging.
Two claims keep getting mixed together. Ghost kitchens can clearly make opening a restaurant much cheaper. They do not automatically turn restaurant operations into a high-margin business.
As this chart shows, and as featured in our ghost kitchen market deck, search interest in ghost kitchens has risen sharply
How much do DoorDash and Uber Eats take from ghost kitchens today?
For a U.S. ghost kitchen buying marketplace delivery today, DoorDash typically charges 15% to 30% of the order subtotal, while Uber Eats now starts at 20% and goes as high as 30%.
Uber's entry price is especially worth updating. Its U.S. Lite marketplace fee increased to 20% earlier this year. Plus costs 25%, rising to 30% on Uber One member orders, while Premium is 30%. Eligible pickup orders cost 7%.
DoorDash currently keeps its U.S. delivery tiers at 15%, 25% and 30%. Pickup is 6%. The expensive plans buy more than transportation: restaurants get wider delivery reach, better positioning and access to frequent DashPass customers.
That is why marketplace dependence can get painful. A restaurant is partly paying for a courier and partly paying for demand.
Take the latest National Restaurant Association median figures only as an illustration. Food, labor and occupancy together represent 69.3% of sales at the median limited-service restaurant. Put a 25% marketplace commission on top and 94.3 cents of every sales dollar has already been allocated. A good ghost kitchen should run below some of those restaurant medians, so this is not a forecast of its actual margin. It shows how little room there is for mediocre execution.
| Cost on $100 of food sales | Example amount |
|---|---|
| Food at limited-service median | $32.40 |
| Labor at limited-service median | $31.70 |
| Occupancy at limited-service median | $5.20 |
| 25% delivery-platform commission | $25.00 |
| Left before packaging, utilities, insurance, marketing and profit | $5.70 |
Is customer acquisition what kills ghost kitchen margins?
Customer acquisition is often a bigger problem than kitchen rent for a new ghost kitchen, because opening a delivery restaurant is easy while getting people to choose it is expensive.
DoorDash now says more than 615,000 restaurants and grocery markets are on its platform. A virtual burger brand can appear on the app within days, but it enters the same marketplace as McDonald's, Chipotle, thousands of established local restaurants and hundreds of other delivery-first concepts.
DoorDash's own merchant data shows how much discovery matters. For new restaurants on the platform in 2025, more than 55% of first orders came from consumers browsing rather than searching for a specific restaurant. That is useful distribution for a new brand, but it also means the restaurant depends heavily on how the marketplace ranks, recommends and promotes it.
The platforms have become much less tolerant of virtual-brand spam. DoorDash currently requires at least 50% of a virtual brand's main menu to differ from other brands at the same address. It also demands a lifetime customer rating of at least 4.0, minimum weekly orders, merchant cancellations below 5%, missing or incorrect orders below 5%, and normally caps one address at ten virtual brands.
A few years ago, one kitchen could create a dozen slightly different names around essentially the same food and hope to occupy more digital shelf space. DoorDash's rules now attack exactly that strategy.
The economics are much better when customers already know what they want. That is why creator brands, existing restaurant chains and genuinely differentiated concepts have become more interesting than anonymous virtual restaurants.
Kitchen capacity is abundant compared with consumer attention.
If you want more recent data on this point, please see our latest ghost kitchen market report.
This chart, included in our ghost kitchen market deck, shows annual VC investment in ghost kitchen startups
Why do multi-brand ghost kitchens make more sense?
Multi-brand ghost kitchens usually have better economics because the same rent, equipment and staff can serve several pools of demand.
Taster is one of the cleanest current examples. CEO Bertrand Peyrat recently said the company had about 140 restaurants, with 127 operating several brands from the same location and only 13 dedicated to a single brand. In other words, roughly nine out of ten Taster restaurants in that network were multi-brand.
Taster reported more than €187 million of systemwide sales last year across nearly seven million orders. That works out to roughly €27 of gross order value per order. The company also said average sales per unit were around €1.4 million and that 70% of its orders arrive at dinner.
That dinner concentration is exactly why shared capacity is valuable. One single-brand kitchen has cooks, fryers, refrigerators and rent sitting behind one demand curve. A kitchen running Pepe Chicken, Starsmash, Spud Bros Express and other concepts has several chances to fill the same capacity during the evening rush.
Wonder pushes the model further. A current Wonder location can produce food from more than 20 restaurant concepts, and one household can order across several of them in a single transaction. One person can get Indian food while another gets pizza without creating two separate deliveries.
Multi-brand operations still have a limit. Adding brands that require completely different ingredients, equipment and cooking processes can create more chaos than extra revenue. The attractive version reuses as much of the kitchen as possible while bringing in genuinely new demand.
So “number of brands per kitchen” is a poor metric on its own. The useful question is how much extra revenue each additional brand brings relative to the extra ingredients, labor and complexity it creates.
Does owning the restaurant brand make a ghost kitchen business better?
Owning the restaurant brand usually makes a ghost kitchen business more valuable because the company controls the reason customers order, rather than just the room where the food gets cooked.
Kitopi's evolution is one of the best examples. The company became famous for operating cloud kitchens on behalf of other restaurant brands. Today, Kitopi explicitly says that description is outdated. It calls itself a hospitality company, operates more than 200 outlets and has built or acquired a portfolio spanning delivery brands, fast casual restaurants and dine-in concepts.
The economics have improved alongside that shift, at least according to the company. CEO Mohamad Ballout said Kitopi served 50 million meals last year, improved EBITDA by roughly ten percentage points and became free-cash-flow positive in the fourth quarter. He also said its core brands grew more than 30%.
Rebel Foods went down a similar path much earlier by owning brands such as Faasos, Behrouz Biryani and Oven Story. Its latest available financials are useful precisely because they stop us from romanticizing the model. Operating revenue grew roughly 14% in FY25 to about ₹1,617 crore and its EBITDA loss narrowed by about a quarter, yet EBITDA was still around negative ₹128 crore and the company remained loss-making.
Rebel also tried to move even closer to the customer with QuickiES, a service promising food from more than 45 brands in under 15 minutes. Inc42 reported earlier this year that Rebel shut QuickiES after roughly a year because of high cash burn.
Wonder is now buying restaurant intellectual property as well. Its latest move was the acquisition of Salt Hank, the viral New York sandwich brand, following other brand acquisitions such as Blue Ribbon Fried Chicken.
Across Kitopi, Rebel and Wonder, the direction is remarkably similar. Owning production capacity gives a company efficiency. Owning a brand gives it demand. The second asset is harder to replace.
If you want more recent data on this point, please see our latest ghost kitchen market report.
This chart, included in our ghost kitchen market deck, shows why Rebel Foods is winning in ghost kitchens
Is licensing virtual restaurant brands actually a good business?
Virtual-brand licensing can be one of the most capital-efficient ghost kitchen models, provided the brand creates enough extra orders to justify another company taking a cut.
Taster currently shows what the more mature version looks like. It builds brands with creators, develops menus and supply chains, and scales them through franchisees and licensees. Instead of financing every kitchen itself, Taster lets local operating partners provide much of the capital and restaurant labor.
The network grew quickly last year. Taster says group sales increased 45%, more than 30 new restaurants opened, and the company reached the number-three position among restaurant groups on French delivery platforms. Two thirds of new openings are now coming from existing partners, according to its CEO, which is a useful sign: operators who already know the economics are choosing to open another site.
The risk appears when brand licensing becomes little more than selling a logo to spare kitchen capacity.
MrBeast Burger became the famous warning. Virtual Dining Concepts was able to put the brand into hundreds of existing restaurant kitchens at extraordinary speed because Jimmy Donaldson already brought a huge audience. The relationship later ended in litigation, with Donaldson alleging that inconsistent food quality was damaging the MrBeast brand and Virtual Dining Concepts disputing his claims.
That conflict exposed the central weakness of distributed virtual restaurants. The customer may be buying because of Person A's reputation while Person B owns the kitchen and Person C employs the cook.
The model works much better when recipes are simple, ingredients overlap with existing operations, training is standardized and the brand owner can enforce quality. Taster's current move toward more physical restaurants also suggests that even a successful virtual-brand network eventually wants more control over how customers experience the brand.
Why are the strongest ghost kitchens opening physical restaurants now?
The strongest ghost-kitchen operators are moving into physical restaurants because pickup, dine-in and visible storefronts can make the same kitchen more valuable.
Taster is unusually explicit about the change. The company was built around digital food brands and delivery, but management now says one of its main priorities is accelerating its physical presence. Taster stores add takeaway and customer visibility to the same multi-brand production model.
Kitopi made that move years earlier and now operates both dine-in and delivery locations. Rebel Foods has built EatSure food courts and physical outlets around several of its brands. Wonder's whole current format combines a shared kitchen with a customer-facing food hall, takeaway and delivery.
The fee structure makes the logic easy to understand. A DoorDash delivery order can cost the restaurant 15% to 30% of the subtotal. DoorDash pickup costs 6%. A customer ordering directly can be cheaper still, even after payment and software expenses.
A physical location also advertises itself. Thousands of people can pass a restaurant sign without the restaurant paying for every impression. An invisible kitchen inside an industrial building gets none of that.
Wonder has taken the hybrid model furthest. It currently lets customers order from more than 20 restaurant concepts in one location, eat there, collect the food or receive delivery. The company has now expanded beyond 147 locations, and its latest rollout into Massachusetts and New Hampshire produced more than 200,000 orders soon after entering Greater Boston.
Taster's latest numbers reinforce the same direction. Its delivery-born network grew 45% last year, yet management is still making omnichannel expansion one of its main priorities.
As with Kitopi, the companies with real scale increasingly want the customer-facing parts of a restaurant back. The industry has ended up somewhere between a conventional restaurant and the original invisible-kitchen idea.
This chart, included in our ghost kitchen market deck, shows annual funding in ghost kitchen startups
Can ghost kitchens get off DoorDash and Uber Eats?
Ghost kitchens can reduce their dependence on DoorDash and Uber Eats, although replacing those platforms means paying for customer acquisition and delivery themselves.
Rebel Foods built EatSure partly to create a direct ordering relationship across its portfolio. Kitopi now has its own consumer app and says its loyalty program has passed 70,000 users in the UAE.
Wonder took a much more aggressive route by acquiring Grubhub for about $650 million. Wonder now controls restaurant concepts, kitchens, a consumer ordering product and a large delivery marketplace. Few food companies have tried to own so much of the transaction.
That vertical integration gives Wonder a chance to keep money that would otherwise leave through third-party commissions, but it also explains the capital intensity. Wonder has raised roughly $3 billion. Its latest financing valued the company at $9 billion before the new money, and the Financial Times currently reports more than 147 operating locations as the company pushes toward further geographic expansion.
DoorDash itself makes clear why going direct is attractive. Its own Commerce Platform now handles digital ordering for more than 150,000 merchants, and revenue from that service grew more than 40% year over year in its latest quarter. Restaurants increasingly want marketplace demand when they need discovery while building cheaper direct channels for repeat customers.
The strongest setup is probably a mix rather than complete independence. A restaurant can use DoorDash or Uber Eats to acquire new customers, pickup and dine-in for lower-cost transactions, and its own app or website for customers who already know the brand.
Trying to replace DoorDash from scratch is expensive. Preventing DoorDash from owning every repeat order is much more realistic.
If you want more recent data on this point, please see our latest ghost kitchen market report.
Can a ghost kitchen work in any city?
A ghost kitchen cannot work equally well everywhere because delivery density changes how much revenue one kitchen can produce and how expensive each order is to fulfill.
Traditional restaurants care intensely about location, and ghost kitchens do too. They simply optimize for a different kind of traffic.
A conventional restaurant might pay more for a corner with thousands of pedestrians. A ghost kitchen wants thousands of potential delivery customers inside a tight radius, easy courier access and enough order density to keep cooks and drivers busy.
CloudKitchens currently concentrates its network in major metropolitan areas and explicitly markets facilities around dense delivery submarkets. Its U.S. footprint includes 12 facilities around Los Angeles, nine in the Bay Area, five around New York and several each in markets such as Chicago and Dallas.
Wonder's current expansion follows similar logic. It built heavy density in the Northeast before moving toward Texas. In New York, Wonder says meals arrive in 35 minutes or less on average. That promise becomes much harder to keep when a kitchen serves customers spread across a large low-density area.
Density also improves multi-brand economics. Twenty restaurant concepts are much more useful inside one kitchen when enough nearby households generate orders across all of them.
Cheap real estate can therefore be misleading. Saving $5,000 a month on rent is a poor trade if the location loses tens of thousands of dollars in potential orders or forces deliveries to travel farther.
Ghost kitchens removed the need for prime pedestrian real estate. They did not remove geography from the restaurant business.
This chart, included in our ghost kitchen market deck, compares the main business model options for ghost kitchen companies
Are ghost kitchens really tech companies?
Ghost kitchens today are restaurant businesses first, with technology becoming valuable when it removes a few points of food, labor or delivery cost.
Kitopi has built a proprietary Smart Kitchen Operating System to coordinate orders, kitchen workload, customer service and other operations. Rebel Foods has spent years standardizing recipes, kitchen processes and supply chains. Wonder is now investing heavily in kitchen automation and robotics.
Those systems can produce a large financial effect because restaurant margins are thin. The latest National Restaurant Association data puts median pretax profit for limited-service restaurants at only 4%. Saving two percentage points of labor or food cost would therefore be huge relative to the typical bottom line.
But software cannot rescue a restaurant that nobody wants to order from.
The industry's own strategic moves are revealing. Kitchen United eventually left physical operations and returned to being a software company. Kitopi went the other way: it kept the technology but became a hospitality group. Taster talks increasingly about creators, franchises and physical restaurants. Wonder keeps buying restaurant brands while building more kitchens.
Technology clearly helps these companies process many concepts through the same operation. It can forecast demand, route orders, reduce mistakes and simplify training. In most cases, the restaurant brand, customer relationship and operating network are still more valuable than the kitchen software itself.
The strongest ghost-kitchen companies these days look more like unusually data-driven restaurant groups than SaaS companies with fryers.
Which ghost kitchen business models actually work today?
The ghost-kitchen models that work today either sell efficient infrastructure to restaurants that already have demand or use shared kitchens to make strong restaurant brands more productive.
The post-pandemic evidence is much clearer now than it was during the boom.
Kitchen infrastructure remains viable. CloudKitchens is still operating at large scale and has increasingly oriented the pitch toward established restaurant chains, where a ghost kitchen can be a cheap way to test or enter another trade area.
Multi-brand operation also has strong evidence behind it. Taster grew systemwide sales 45% last year with almost its entire network operating several brands per kitchen. Wonder now puts more than 20 concepts under one roof. Kitopi says it served 50 million meals last year and became free-cash-flow positive in the final quarter after years of moving deeper into brand ownership and omnichannel restaurants.
Virtual-brand franchising can work particularly well because the brand company avoids much of the restaurant capital expenditure. Its weakness is quality control, especially when the brand owner and kitchen operator are different companies.
The models to be most skeptical of are anonymous delivery-only concepts that rent a kitchen, have no existing audience and depend on marketplaces for almost every order. That business can save money on rent and still lose the economics through food cost, labor, platform commissions, promotions and low kitchen utilization.
The broader industry has already voted with its capital. Kitchen United abandoned physical kitchens. REEF sharply reduced its kitchen ambitions. Rebel closed a cash-hungry 15-minute delivery experiment. Meanwhile Taster is adding physical stores, Kitopi has become a restaurant group, Wonder keeps buying brands and CloudKitchens increasingly pitches established chains.
The original ghost-kitchen idea focused too heavily on removing the dining room. Today the value is easier to see: a brand people deliberately choose, enough local demand to keep the kitchen busy, several ways to reach the customer and an operation that can produce more orders without adding costs at the same rate.
A ghost kitchen is best understood as restaurant infrastructure rather than a restaurant strategy on its own. Put a strong brand and dense demand into that infrastructure and the economics can be excellent. Put an unknown brand into it and pay a marketplace to supply every customer, and removing the dining room does surprisingly little.
| Ghost kitchen business model | Our judgment today | Why it can work | What usually breaks it |
|---|---|---|---|
| Multi-brand operator with strong owned brands | Strong | Several demand pools share one kitchen | Operational complexity |
| Virtual-brand franchise or licensing network | Strong when the brand creates demand | Low capital needs and fast expansion | Inconsistent food quality |
| Kitchen rental for established restaurant brands | Viable | Cheap, fast geographic expansion | Poor occupancy |
| Hybrid food hall with delivery, pickup and dine-in | Increasingly convincing | More channels improve kitchen utilization and customer acquisition | Higher capital cost |
| Fully integrated food and delivery platform | Potentially powerful, still expensive | Captures more of the transaction | Huge capital and logistics burden |
| Unknown delivery-only brand dependent on marketplaces | Weakest | Very easy and cheap to launch | Platform fees, weak demand and low differentiation |
If you want more recent data on this point, please see our latest ghost kitchen market report.
This chart, featured in our ghost kitchen market deck, shows how revenue is distributed across customer segments in the ghost kitchen market
OUR METHODOLOGY
This analysis tests which ghost-kitchen business models appear durable today by separating businesses that are often grouped under the same label. We compare shared-kitchen rental, owned multi-brand restaurant operations, virtual-brand franchising, managed kitchens and vertically integrated food platforms rather than treating “ghost kitchens” as one economic model.
For each model, we looked at the operating factors that determine whether the economics hold together: delivery demand, kitchen utilization, restaurant cost structures, marketplace commissions, customer acquisition, geographic density, capital requirements, brand ownership, direct customer relationships and evidence of profitable or cash-generative operations.
We gave more weight to operating evidence than to funding rounds or headline valuations. Closures, retrenchments, repeat franchise openings, sales per unit, free-cash-flow disclosures, restaurant-level cost data and changes in channel strategy are more useful here because they show what businesses are actually keeping, abandoning or scaling after the pandemic-era boom.
Company-reported figures are used to understand scale and strategic direction, but promotional claims are not treated as independent industry benchmarks. CloudKitchens' sales and opening-cost claims, for example, help explain its current pitch to established restaurant brands, but they are not used as a universal measure of ghost-kitchen performance.
Restaurant operating-cost benchmarks come primarily from the National Restaurant Association's 2025 data, including food, labor, occupancy and pretax-profit ratios for limited-service restaurants. Current marketplace economics are based on DoorDash's U.S. merchant pricing and Uber Eats' 2026 U.S. marketplace terms.
We also compared post-pandemic expansion and contraction decisions across major operators. Kitchen United and REEF are useful failure or retrenchment cases; Taster, Kitopi, Rebel Foods, CloudKitchens and Wonder provide different examples of multi-brand operations, licensing, brand ownership, kitchen rental, omnichannel expansion and vertical integration.
Key sources used for this analysis include: DoorDash's Q2 2026 financial results, the National Restaurant Association's Off-Premises Restaurant Trends 2025, the National Restaurant Association's 2025 Restaurant Operations Data Abstract, DoorDash's current U.S. merchant pricing, Uber Eats' U.S. merchant terms, CloudKitchens' current operating disclosures, Restaurant Business on Kitchen United's shutdown and software pivot, Zepros' interview with Taster CEO Bertrand Peyrat, Mohamad Ballout's Kitopi operating update, Inc42 on Rebel Foods' FY25 financial performance, and the Financial Times on Wonder's scale, capital base and vertical-integration strategy.
This chart, included in our ghost kitchen market deck, shows how multi-brand kitchen management technology has evolved over time
Related blog posts
