Latest / Restaurant Business with Fexingo: Food Service, Hospitality, and Independent Operators / Why Restaurant Ghost Kitchens Are Failing in 2026
Transcript
- Lucas: You remember the ghost kitchen hype, right? Around 2020, 2021 — every VC was throwing money at the idea of restaurant kitchens with no dining room, no host stand, just a hood system and a tablet pinging with DoorDash orders. Luna: Oh, I remember. I toured one in Chicago back in 2022. Looked like a warehouse full of induction burners. The guide told me they could 'scale infinitely.' I remember thinking, who eats at a place with no windows? Lucas: That's exactly the question that's catching up with the industry in 2026. Because the numbers are not pretty. A firm called Foodservice Analytics tracks about 2,400 ghost kitchen facilities in the US right now. They estimate that roughly 40 percent of them are operating at less than 50 percent of their capacity. Luna: Half empty. That's brutal when your rent is fixed. Lucas: And the rent is not cheap. A lot of these operators signed leases in 2022 and 2023 at peak industrial real estate prices. They were competing with last-mile logistics warehouses for the same spaces. So you've got a $12,000-a-month kitchen with a single delivery brand making maybe $30,000 in monthly revenue. After food cost, labor, and platform commissions, you're in the red. Luna: Let's talk about those platform commissions. DoorDash and Uber Eats are taking somewhere around 25 to 30 percent per order, depending on the plan. For a ghost kitchen with no dine-in revenue to offset that, the margin math is terrible. Lucas: It's the core problem. A traditional restaurant can subsidize its delivery channel with in-house diners who buy the higher-margin drinks and appetizers. A ghost kitchen has only delivery. So every order bears the full commission weight. And the customer is already paying a higher price on the app because the operator has to mark up the menu to cover the fee. Luna: So you're charging $18 for a burger that costs you $6 to make, and after the platform takes its cut, you're left with about $7. Out of that, you pay rent, utilities, labor, insurance. It doesn't take long to figure out you need to sell a lot of burgers. Lucas: And most ghost kitchens aren't selling enough. The average ghost kitchen does about 80 to 120 orders a day. A busy McDonald's does more than triple that out of a single drive-through. The throughput just isn't there. Luna: Let's look at a specific case. CloudFoods — they raised $450 million, built 30 locations in New York and Chicago. They were supposed to be the WeWork of restaurant kitchens. What's their status now? Lucas: CloudFoods is still alive, but just barely. As of June 2026, they're running at about 55 percent capacity across their portfolio. They've closed four locations already. The problem is they sublease kitchen slots to individual virtual brands, and those brands are failing. When a brand like 'Burger Republic' or 'Wing Dynasty' shuts down after six months, CloudFoods has to find a replacement fast or eat the vacancy cost. Luna: And the replacement is often another brand with no track record. It's a churn machine. Lucas: Exactly. The failure rate for virtual brands launched inside ghost kitchens is something like 70 percent within the first year. Compare that to the roughly 20 percent annual failure rate for traditional independent restaurants. The ghost kitchen model was supposed to lower the barrier to entry, but it turns out that barrier exists for a reason. Luna: There is one operator that seems to be doing it right: Sweetgreen. They have a bunch of off-premise 'Sweetgreen Outposts' in urban areas. But they're not a pure ghost kitchen — they use those locations for delivery and pickup, but they also have their own app, their own loyalty program, and they're not relying on DoorDash for 90 percent of orders. Lucas: Right. Sweetgreen controls the customer relationship. That is the critical distinction. Most ghost kitchen brands are invisible to the customer. You order 'Freddy's Famous Tacos' on Uber Eats, but you have no idea where it's coming from. You have no loyalty to that brand. Next week, if the kitchen switches to 'Sally's Sushi Bowls,' you'll order that instead. The brand equity is zero. Luna: chick fil a is another interesting case. They've been testing takeout-only locations in the suburbs. But again, they're chick fil a. They have decades of brand trust. And they're not paying 30 percent commission because they handle their own digital ordering. Lucas: Which brings us to the core lesson for independent operators: if you're thinking about a delivery-only pivot, you need to own the channel. That means investing in your own website, your own app, maybe a text-message ordering system. You can't hand your entire economics to DoorDash. Luna: But that's expensive. Building an app and driving downloads is not cheap. Lucas: It is. But the alternative is running on negative margins. There's a middle path: use DoorDash for discovery, but push every repeat customer onto your own system. Put a card in every bag that says 'Order direct next time and save 10 percent.' Even if only 15 percent of customers switch, that fifteen percent becomes your profitable base. Luna: That's smart. It's the same strategy that coffee shops use: use the Starbucks app to find the store, then join the local loyalty program. Lucas: Right. And independent operators have one advantage that ghost kitchens don't: a physical location. A storefront builds trust, it builds habit, it builds a neighborhood presence. When you're a ghost kitchen, you're just a line on a screen. That's fragile. Luna: Speaking of fragile — I want to bring up something that's been on my mind. We've been doing this show for nearly a hundred episodes now, and it's entirely listener-supported. No ads, no sponsors trying to sell you the latest POS system. Lucas: Yeah, and that's only possible because a small group of listeners chip in monthly through Buy Me a Coffee. It's at buy me a coffee dot com slash fexingo. Literally just a handful of people covering the hosting costs and a bit of research time. It means we never have to spin an episode to please a sponsor. Luna: Exactly. And we want to keep it that way. So if these conversations have helped you think differently about your restaurant's numbers, that's the way to keep it going. No pressure — just putting it out there. Lucas: Alright. Back to the economics. Let's talk about what happens when a ghost kitchen fails. A company called Kitchen United, which was backed by Kroger and Google, filed for bankruptcy in early 2025. They had 20 locations and burned through $300 million. Luna: And that's not the only one. I've seen reports that several smaller regional operators have just locked the doors and walked away. The equipment gets auctioned, the lease goes dark. Lucas: There's a secondary market now for used ghost kitchen equipment — hood systems, prep tables, walk-in coolers. It's a sign that the shakeout is real. But I don't think the concept is dead. I think it's just being forced to mature. Luna: What does 'mature' look like? A ghost kitchen that works? Lucas: I think it looks like a multi-brand operator with very high volume in a dense urban area, using its own delivery network — think of a cloud kitchen that employs its own drivers for a two-mile radius. That cuts the commission from 30 percent to maybe 10 percent for the cost of a few drivers. Luna: You would need a lot of orders to justify that fleet. But in Manhattan or downtown San Francisco, maybe it works. Lucas: Exactly. The model only works at density. A ghost kitchen in a suburban strip mall with a population density of 3,000 per square mile? That's a recipe for failure. You need at least 15,000 per square mile to get the order density to make the unit economics positive. Luna: So the real lesson might be: location still matters. Even when there's no dining room. Lucas: Absolutely. And the second lesson is: brand matters. The most successful delivery brands are the ones that already exist in the physical world. Domino's, McDonald's, Chipotle — they do massive delivery volume because customers already know exactly what they're getting. A ghost kitchen starting from zero has to build that trust from scratch, and that's expensive. Luna: So what's your advice to an independent operator listening right now? Someone with a single location who's thinking about adding a delivery-only brand on the side? Lucas: I'd say: don't do it as a standalone. If you already have a kitchen and some slack capacity, you can test a virtual brand using your existing equipment and staff during off-peak hours. That's what many pizzerias do — they sell 'artisan flatbreads' on DoorDash at 2 PM. The incremental cost is low. But signing a separate lease for a ghost kitchen? I'd run the numbers very carefully. The break-even is usually around 200 orders a day, and most operators never get there. Luna: So the ghost kitchen dream is not dead — but it's been downsized from 'disrupt the entire industry' to 'a tactical option for specific operators in specific neighborhoods.' Lucas: Exactly. The hype cycle is over. What's left is a real business, but a much smaller one than the PowerPoints promised. And that's fine. The restaurant industry doesn't need to be disrupted. It just needs to make a profit.