The restaurant industry will likely remember 2023 as the year ghost kitchens died.
After years of hype and millions invested, the pioneers in the delivery-first, commissary-style shared kitchen facilities have either disappeared, pivoted to software or food halls, or show signs of ill health.
Reef Kitchens, for example, once had a deal with Wendy’s to open up to 700 delivery-only locations out of mobile kitchens in parking lots, but the fast-food chain announced plans to significantly scale back its use of Reef Kitchens in 2022. Then, in March, Reef started winding down its ghost kitchens and instead licensed its technology to venues like airports and stadiums.
Roughly a year after receiving a $100 million investment from a group that included grocery giant Kroger, Kitchen United in November suddenly announced it would be closing all eight of its outlets in the grocery chain and would sell or close its other brick-and-mortar locations. The company instead would pivot to selling software.
C3, which had promised to build a network of more than 200 ghost kitchen locations across the country, has instead transformed its brick-and-mortar locations to food halls offering “immersive, in-person dining experiences,” according to a press release.
CloudKitchens, the startup by Uber founder Travis Kalanick that arguably gave birth to the ghost kitchen trend, in September was reportedly laying off staff and closing locations. (Kalanick is also reportedly now backing a company building automated makelines.)
And yet those in the space insist that rumors of the ghost kitchen trend’s demise have been greatly exaggerated.
Are ghost kitchens really dead?
That depends on what you call a ghost kitchen.
To be clear, the death knell has focused specifically on the commissary-style ghost kitchen facilities, with multiple kitchens rented by restaurant operators to support delivery, mostly, with little-to-no dine in.
“That model sucks. That model is horrible,” said Gregg Majewski, founder and CEO of Craveworthy Brands, which is growing its own family of delivery-only brands. “It’s not financially viable for anybody, renting a little kitchen. You can’t do the volume out of those.”
But out of the ashes of the first wave of ghost kitchens come a variety of new models positioned for growth. These are concepts that attempt to capitalize on some of the features of ghost kitchens that still show promise.
There’s ClusterTruck, for example, which has five locations and plans to launch franchising this year. In New York, Nimbus has taken over two former Kitchen United facilities and now has four shared-kitchen facilities.
In Texas, Oomi Digital Kitchen is plotting expansion with a hub-and-spoke model. Atlanta-based Experiential Kitchen is also franchising and has three brick-and-mortar locations with two more in the works. In Reno, Nev., Vistro combines multiple brands in a dine-in/takeout/delivery setting.
They vary in their approach, but these are mostly delivery-focused concepts with multiple brands, sometimes playing host to other operators.
These 2.0 models do have one thing in common, however: They don’t call themselves ghost kitchens. That’s a term many hope will fade into the mist.
“You have to burn the ships,” said Chris Baggott, chairman of both ClusterTruck and the software spinoff Empower Delivery, which is considered part of the new wave.
No one wants to be associated with the “$3 to $4 billion that disappeared last year in those failed experiments,” he said.
What went wrong?
When ghost kitchens first entered the scene around 2019, they were designed to solve a lot of problems for restaurant operators.
The idea was that the shared kitchens offered a lower-cost space in which young brands could find their footing, or perhaps more established brands could expand their reach into new delivery markets without the capital investment of brick-and-mortar real estate.
And when the pandemic hit, finding a model that allowed for low-cost expansion with delivery suddenly seemed like a very, very good idea.
Facilities like CloudKitchens initially went into large warehouse-like spaces off the beaten path, where delivery drivers could find them, but consumers were not initially welcome. The CloudKitchens brand, to this day, is not really known to consumers. Most are called something like “Costa Mesa Kitchens,” implying more of a food hall setting.
Reef was designed to operate out of trailers, or “vessels,” in parking lots as “proximity hubs,” also accessible to delivery drivers but not consumers.
Kitchen United and C3’s Citizens concept later included on-site ordering and pickup features. But fundamentally these were models that involved an intense capital investment, said Meredith Sandland, who was chief operating officer of Kitchen United from 2018 to 2020 and now is CEO of the software company Empower Delivery.
The ghost kitchen pioneers “were very reliant on many different participants to make it work,” Sandland said.
“You had an underlying landlord, you had an operator, you had a restaurant brand that maybe had a franchisee, you had the third-party delivery companies,” said Sandland. “There were so many people trying to split the profit.”
The early ghost kitchens then added pickup, in part because they couldn’t make the economics of delivery-only work, she said. “And, as they incorporated pickup, they became more and more like restaurants and less and less like software companies.”
That doesn’t mean the early ghost kitchen model is terrible and can’t work, she added. “It just means there was a misalignment between venture-capital expectations and what they were trying to do.”
From ghost kitchens to 'host kitchens'
Rishi Nigam, CEO of Atlanta-based virtual brand matchmaker of sorts Franklin Junction, said he never thought the shared-commissary ghost kitchen model was going to work—even though those facilities were sometimes clients of Franklin Junction.
Nigam’s company helps restaurants do more out of underutilized kitchens by adding virtual brands for delivery. Those restaurants become “host kitchens,” a term Franklin Junction has trademarked.
Denny’s, for example, recently announced a partnership with Franklin Junction in which the family dining chain’s restaurants will serve as host kitchens for as-yet-unnamed virtual brands for delivery only, bringing incremental revenue during off-peak hours. Franklin Junction helps those host kitchens find the right virtual brands to operate, and helps the virtual brands find the right host kitchens—a matchmaker.
Over the past few years, ghost kitchen operators have come to Franklin Junction to find brands to add capacity to their kitchens too.
But Nigam said those early pioneers faced a number of challenges: moving to class B or C real estate, for example, and oversaturating markets with delivery brands competing for the same audience.
“We never thought that was going to be a good model,” he said.
Now, however, Nigam is seeing the new wave of operators knock on his door. These, he contends, have potential.
Generally these tend to be smaller—“mini ghost kitchens”— run by single operators who can “control quality, curate food for the neighborhood and not be so hidden,” he said.
Here’s a look at a few examples in the ghost kitchen 2.0 landscape:
ClusterTruck offers 15 in-house brands for delivery only. | Photo courtesy of ClusterTruck.
Not a truck
The amusingly named ClusterTruck is meant to evoke the gathering of food trucks that give consumers a chance to experience multiple menus in one setting, but it was never a truck.
Dubbed a “delivery-only kitchen,” the concept was launched in 2016 in Indianapolis. Now there are three in that market, as well as locations in Kansas City, Mo., and Columbus, Ohio.
The company is also launching franchising this year. And a spinoff of ClusterTruck is the software company Empower Delivery, headed by Sandland, which makes the tech infrastructure available to others who want to build their own virtual brand ecosystem.
ClusterTruck locations make 15 brands available for delivery out of one kitchen, and all of the brands are created in house. Delivery is also mostly operated in house.
“The only way you can do it is to be vertically integrated,” said Baggott. “You have to control everything.”
Consumers are ordering food from ClusterTruck, not the individual brands. Baggott thinks of ClusterTruck kitchens as operating like a Cheesecake Factory, a brand long known for its massive menu. The 15 brands under the ClusterTruck name are like menu subcategories, including about 90 items.
“The sauté station doesn’t care if its Chinese or Mexican or Indian,” he said.
ClusterTruck’s software is designed to optimize the timing of delivery. When an order comes in, the software recognizes how far out the (in-house) driver is and knows that the pad Thai, for example, needs six minutes, or a cheeseburger might need four, ensuring that the food is delivered fast and hot.
Sure, ClusterTruck exists on third-party delivery marketplaces too, but those platforms only accounts for less than 6% of volume, said Baggott. Consumers who order through the ClusterTruck app or website pay no delivery or service fees.
One thing consumers cannot do at a ClusterTruck location, however, is pick up their food.
“We don’t like pickup. I hate pickup,” said Baggott. Before Kitchen United struck a deal with Kroger to open outlets in grocery stores, ClusterTruck tried it. Kroger wanted customers to be able to pick up.
The attempt was hampered by Covid and supply chain issues, Baggott said, but, in the end, it didn’t work.
“I told them, don’t do this,” said Baggott. “Pickup messes with your delivery system. We designed a system to only do delivery.”
Oomi Digital Kitchen in Dallas uses the ClusterTruck software. | Photo courtesy of Oomi Digital Kitchen.
Compatible by design
One example of a company using ClusterTruck’s Empower Delivery software is Oomi Digital Kitchen out of Dallas.
Co-founder Markus Pineyro calls his model a delivery kitchen. Like ClusterTruck, Oomi is vertically integrated, controlling everything from the ordering platform to delivery out of one central kitchen offering six in-house brands.
“The key is there are no licensed brands,” said Pineyro. “It allows us to cross utilize ingredients, staff, equipment. It brings efficiencies that are not usually seen.”
The same guy who’s grilling chicken for tacos is grilling chicken for a salad, for example, or the person who’s frying egg rolls might also be frying chicken tenders.
But unlike ClusterTruck, the 2,200-square-foot Oomi location welcomes customers who prefer to pick up their food—though there is no dine-in option.
“Our kitchen is centrally located on a main street close to downtown,” he said. “It’s an open kitchen and people can see in through the windows. It’s very brand forward. They know who we are. We’re not in a warehouse district hidden in the back.”
Also unlike ClusterTruck, Oomi does about 55% of sales through third-party delivery. The rest is handled by in-house drivers who are kept to a tight radius. Either way, consumers are charged a flat $2 delivery fee.
Oomi has only been open a little more than a year, but Pineyro envisions expansion with the central kitchen serving smaller delivery outlets in a hub-and-spoke system.
“We’re looking to add brands without adding complexity,” he said.
Nimbus picked up two Kitchen United locations in New York City. | Photo courtesy of Nimbus.
A co-cooking company
Nimbus founder and CEO Camilla Opperman believes ghost kitchens can work. “They just need to be executed correctly,” she said.
The closure of Kitchen United locations in New York City last year enabled the then two-unit Nimbus to double in size by picking up the facilities.
Opperman described Nimbus as a “shared kitchen, co-cooking company” that combines hourly and long-term kitchen rentals. Clients can cook there for a few hours, or for a few years, and Nimbus works with a variety of businesses, from caterers and bakers, to CPG brands and popups.
“Most of our competitors were focused on serving delivery businesses, but we have a wider range of people using our space,” she said.
Facilities range from 7,000-square-feet to 10,000-square feet and host seven to 14 rentable units, which are a mix of closed-door venues and more open kitchens or dedicated stations where brands can have their specific equipment.
“Ultimately the goal is to increase transparency,” said Opperman. “We want people to know their food is coming from a clean, safe space.”
In some ways, Nimbus is like a traditional ghost kitchen. Jersey Mike’s, for example, uses the facility to expand its delivery reach. Even DoorDash is a customer, offering outlets for some of its local hero brands, she said.
Opperman, however, argued that it was the vertical integration that killed the early ghost kitchens.
Operating kitchen spaces, developing technology to enable delivery and launching virtual brands are all very different businesses, she contended. “That’s way too complex.”
Nimbus, meanwhile, is staying focused on the kitchen infrastructure “and operating those as seamlessly as we can,” she said.
A hybrid model
Out of Atlanta comes Experiential Brands, launched by private-equity firm NRD Capital in 2021.
Experiential Brands franchises both a brick-and-mortar model or delivery-only brands. A corporate location in Sandy Springs, Ga., is like a mini food hall, with the company’s four in-house brands: The Original Hot Chicken, Inked Tacos, Pinsa Roman Pizza and Flametown Burgers. All are also offered for delivery.
Franchised brick-and-mortar locations in Woodstock, Ga., and Houston house two of the brands.
Another franchise unit is coming to Toronto, and another corporate location is planned for Southern California, though those will likely start out with only one of the brands because they are in higher-volume markets.
A priority is providing fantastic customer service “whether that customer is standing in front of us or we’re delivering to their home or office,” said James Walker, Experiential Brands’ chief concept and culinary officer.
The company’s portfolio was developed to serve the goal of founder Aziz Hashim, which is returning profitability to franchisees, he said. The four brands were developed to be served in brick-and-mortar locations, but they have been pulled into the virtual brand world because of demand.
Frisch’s restaurants in Ohio, for example, are serving as host kitchens offering the Original Hot Chicken and Flametown Burger brands for delivery only. It’s another match made by Franklin Junction.
There is also much opportunity for ghost kitchens overseas, Walker contends. The ghost kitchen provider Kitopi, for example, will be operating the Experiential Brands portfolio in the Middle East under a licensing agreement.
The world of delivery kitchens, as Walker calls them, is evolving.
“I think it’s evolving to platforms like Franklin Junction and their host kitchen business model,” he added. “You’re seeing some of the businesses move from [the failed ghost kitchens] to host kitchens.”
Vistro in Reno, Nev., offers delivery, pickup and dine in with a mix of in-house and licensed brands. | Photo courtesy of Vistro.
The virtual bistro
In Reno, Nev., a “delivery-first kitchen” is planning to add Experiential Brands’ Original Hot Chicken to their brand mix—yet another match made by Franklin Junction.
Dubbed Vistro—for “virtual bistro”—the concept offers multiple menus out of one kitchen, including eight in-house brands. The facility serves as a host kitchen for three licensed brands: Bennigan’s, Nathan’s Famous Hot Dogs and Tori-Yoshi (a Franklin Junction-created brand) with Original Hot Chicken scheduled to start next week.
Unlike the others, Vistro offers delivery, takeout and dine-in.
Derek Sornson, owner-operator of Vistro contends it is the delivery-only focus that killed the early ghost kitchens.
Vistro first opened as a delivery-only concept available through third-party marketplaces. “It was successful in terms of gross sales, but not in terms of gross profit, because of the fees,” he said.
Then Vistro moved to a prominent location with a bright orange building to catch attention. There’s indoor and patio seating, a liquor license and event space. Reno is a 24/7 town, so there’s a brisk late-night business.
Vistro is still available on third-party marketplaces, but the business doesn’t have to depend solely on that revenue stream.
“What we learned is that, yeah, delivery works, but too bad there are so many fees associated with everything. And not everyone is completely ready to order everything at home,” said Sornson. “They still want that in-person experience.”