Progress in restaurant robotics hit a setback late last month.
Kernel, the robot-powered vegan concept created by Chipotle founder Steve Ells, is no more.
As my colleague Lisa Jennings reported, the two-unit New York-based restaurant is rebranding as Counter Service, a chef-driven sandwich concept with no robots.
Though Kernel was small, it was ambitious. It believed that by automating some elements of kitchen work, the restaurant could operate with fewer employees while paying them more. Ells told the New York Times in October that Kernel was building “the platform of the future” for restaurants.
But now its robot is gone, and with it goes another potential blueprint for the automated restaurant.
It’s not entirely clear what went wrong. COO Tom Cortese told Jennings that Kernel’s plant-focused menu didn’t have broad appeal, although people who tried it left rave reviews.
But he also suggested that the company wanted to be less robot-centric. Counter Service will still use a lot of technology, but it will mainly be software that works in the background.
The Times story hinted at some technical difficulties with Kernel’s robot. Late last year, the company shut down its restaurants temporarily to revamp the bot. Ells told the paper that automating restaurants is a “wickedly complicated endeavor.” Just a few months later, he has moved on to something else.
The story of Kernel is a good case study for why restaurant robots haven’t gone mainstream yet. Robots may still be unreliable, or they may not be a good fit operationally, even when they’re tailormade, as Kernel’s was. As much as we like to think of kitchens as mini food factories, cooking is ultimately a lot different than building car parts or moving pallets around a warehouse. Food is fluid and delicate and on-demand. Like Ells said, it’s complicated.
On top of that, there’s the consumer piece. Even in the AI era, customers are still not crazy about the idea of a robot making their food. As of last year, just 30% of adults said they’d like to order food that gets prepared by a robot, according to the National Restaurant Association’s Restaurant Technology Landscape Report. That means 7 in 10 might have turned away from a place like Kernel.
You put those two things together and you start to get an idea of why the concept failed. (There was also the whole vegan thing.)
That’s not to say robots can’t work in restaurants. Just look at Sweetgreen. The salad chain has had some early success with its so-called Infinite Kitchen—a conveyor belt that dispenses ingredients into bowls as they move down the line.
The robots, now in 12 locations, have lowered the chain’s labor and food costs and have boosted sales by moving customers through the line faster. Ninety percent of customers that have visited a robot location said it was a positive experience. And half of the 40 restaurants the chain plans to open this year will have a robot in them.
“We continue to see proof points of operational efficiency, throughput and improved guest experience from the Infinite Kitchen,” CEO Jonathan Neman said during an earnings call last month.
What did Sweetgreen do that Kernel couldn’t? For one thing, it had a head start. Sweetgreen was already a large, well-known brand with a popular product when it began installing robots. It also bought a robot that had already been functioning for years in a pair of restaurants in Boston.
Kernel was starting from scratch and was immediately labeled a robot restaurant, not to mention a possible Restaurant of the Future. Its fate therefore would hinge on that robot. And Ells' pedigree meant the whole world would be watching. One wonders if, had Kernel had a bit more time and space to tinker, it could have made it work.
Sweetgreen's menu also may be more of a natural fit for automation. Almost everything the chain sells comes in a bowl. Kernel sold salads, but also sandwiches and burgers, roasted vegetables and other sides. Its robot system involved an arm, conveyor belts and automated ovens. Most of the food was prepped off-site, so it was essentially just heating things up. But still, there were a lot of moving parts, which meant a lot could go wrong.
That’s not to take anything away from what Sweetgreen has done with the Infinite Kitchen so far. If it keeps this up, it could end up being the restaurant-robot blueprint that Kernel dreamed of.