Uber founder Travis Kalanick’s new tech company emerged from “stealth mode” last week after eight years of operating mostly in secret.
The company, previously known by the generic name City Storage Systems, is now called Atoms. It encompasses Kalanick’s existing restaurant-related projects, Otter POS, Lab37 food robots, the Picnic office catering service and, of course, CloudKitchens, a network of ghost kitchens with locations in 30 countries.
And, because all of that is apparently not complicated enough, Atoms also has two other verticals focused on transportation and mining.
The goal of all three businesses is automation—of restaurants, cars and mines—using “gainfully employed” robots, i.e., robots that are designed to do one job rather than many.
Atoms believes this robotic triforce will help bring about nothing less than a new Golden Age for civilization—a world where growing, making and moving physical things can be accomplished without human labor.
In other words, it could not be more ambitious.
In a lengthy mission statement on Atoms’ website, though, Kalanick acknowledges that we’re “less than a millionth of the way there.” And details about how it plans to get there are scarce. (The company has not responded to multiple requests for more information.)
That means for now, the buzz about Atoms is mostly tied to Kalanick’s past accomplishments.
Kalanick is a controversial figure. He built one of the world’s most consequential and disruptive companies with Uber, becoming Silicon Valley royalty in the process. He was also accused of creating a culture of toxic masculinity at Uber and allowing harassment and discrimination to go unchecked.
In his Atoms’ manifesto, Kalanick suggests that he was unfairly forced out of Uber, and he seems to have something to prove with his re-emergence.
When a person like that turns their attention to restaurants, it demands attention, and scrutiny.
We’ve dug into CloudKitchens extensively and have found some disturbing patterns and practices (a bit more on that later). But CloudKitchens is just one piece of Kalanick’s larger plans for restaurants. So let’s take a step back and lay out everything we know about what that could look like.
Kalanick has variously described his vision for the food part of Atoms as “infrastructure for better food,” “automated burritos” and a “food computer.” He has said he wants to do for food what Uber did for cars, and that he wants to make restaurant delivery as affordable as buying groceries.
(Some of this makes sense, as far as elevator pitches go. But then there’s stuff like “food computer,” which seems geared more toward the tech crowd than Atoms’ actual audience, which, presumably, is restaurant operators. It makes me wonder: Who is all of this for, exactly?)
The problem with restaurants, Kalanick says, is that they are like factories that never modernized. While manufacturing in other industries has become increasingly automated, restaurants still rely largely on human labor and all the costs and complications that entails. Labor typically makes up about 30% of a restaurant’s total expenses.
“If you look at the labor statistics, etc., restaurants fit under manufacturing for obvious reasons,” Kalanick said during an interview on the tech podcast TBPN earlier this month. “It just hasn't changed in 50 years.”
(Let’s set aside for now whether this characterization is accurate and whether we really want to live in a world where restaurants become food factories. These are topics for another column.)
Atoms’ first effort at a solution to this perceived problem is Bowl Builder, an automated assembly line that can churn out anything that goes in a bowl, such as rice, protein, grains and veggies. (No sticky or delicate substances or long noodles allowed, and yes, someone still has to prep all of the ingredients.)
According to its website, Bowl Builder can produce 300 bowls an hour, or 200 if they’re being robotically bagged, while reducing marginal labor costs by 30% to 50%, which equates to about two workers on a five-person staff. Elsewhere, the site says Bowl Builder requires just one or two employees per shift.
Though we don’t know the cost of the machine, it is likely going to require a significant investment for any restaurant. It is bulky and L-shaped and has multiple components, including a conveyor belt, a bagging system and a machine that puts lids on containers. It measures 60 square feet and though installation is said to take just one or two days, it would seem to call for major changes to a kitchen’s layout, not to mention operations.
Lab37's Bowl Builder. | Photo courtesy of Lab37
Bowl Builder was developed by Lab37, a team that has its roots at Carnegie Mellon’s Robot Institute. Over the years, the group developed a number of military, mining and agriculture robots before moving to Uber, where it worked on self-driving cars. It has been working on Bowl Builder since 2021.
It’s far from the only company trying to automate bowls. There’s Spyce, the bowl maker acquired by Sweetgreen and recently sold to Wonder, whose plans for restaurants are quite similar to Atoms’.
There’s also Hyphen, the automated makeline being tested by Chipotle.
So far, they have not had the level of impact that Bowl Builder claims it can offer.
Spyce, which Sweetgreen calls Infinite Kitchen, is in about 20 locations today and has reduced labor costs in those stores by 7% to 8% compared to its traditional locations. It is also saving about 1 point on food costs by dispensing more precise portions.
Chipotle has not given specific numbers on Hyphen, but said in 2023 that it did not expect the makeline or other robots, like the Autocado machine, to get it back to a 23% labor cost margin, which would require deflation of 8%. Rather, it believes the bots can help restaurants leverage their labor by allowing them to fulfill more orders.
It’s also worth noting that both Sweetgreen and Chipotle are struggling with significant sales and traffic declines right now. Some have suggested that consumers are losing their appetite for “slop bowls,” especially as those bowls have become more expensive. Now how about slop bowls made by a robot?
Of course, the idea is that Bowl Builder’s bowls will cost less and be more consistent due to the whole no-human-labor thing, giving it an edge on the competition.
As of now, that doesn’t really appear to be happening.
Lab37 created a handful of restaurant concepts to test Bowl Builder. There’s Hungry Cowgirl, Meat + Rice and Farmstand, all comparable to Chipotle, as well as Pita Dust, a Cava analog.
All are offered out of the Mill Food Co., a CloudKitchens facility in Lab37’s hometown of Pittsburgh. (The brands also have outposts in CloudKitchens locations in LA and New York.)
At Hungry Cowgirl in Pittsburgh, the Basic Cowgirl Bowl starts at $10.99. And it is truly basic. Customers can pick a base, protein and sauce, while anything else, like a side of guac ($4.50), costs extra. The bowl also apparently comes with beans and cheese, but there is no option within the online ordering flow to add or remove those.
At Meat + Rice, the opening price point is $9.99 for a base, protein and sauce, while Pita Dust starts at $11.99 for a similar construction, plus cucumbers and feta.
Add to that the $20 minimum for delivery orders, the delivery fee and the tip, and the simplest possible meal from Lab37’s robotic restaurant concepts comes out to more than $30. Not bad for delivery, but not ground-breaking, either. It is comparable to the cost of a similar meal from Chipotle in the same market, where a chicken burrito bowl starts at $11.90.
To be fair, these brands could simply be tests intended to provide a proof of concept for prospective customers. But the whole system still seems to be a long way from making the sort of impact on food delivery that Kalanick envisions.
That could come later. Atoms wants to build robotic cars that could handle the delivery of those bowls, further lowering costs. And it believes it can create network effects with its dozens if not hundreds of ghost kitchens, each of which houses multiple restaurants, including some that could be produced with the Bowl Builder. The fast-growing Picnic catering marketplace, meanwhile, could help generate large sales volume from office buildings, at least around lunchtime.
We could give Atoms the benefit of the doubt on all of that, but the company’s track record gives us pause.
As we have reported, CloudKitchens has been a nightmarish experience for many of the small restaurants that signed up believing it would help them grow their business. We have talked to operators who invested their life savings into opening a delivery-only kitchen only to quickly pull the plug because they were losing so much money.
Part of the issue is that the ghost kitchen model is flawed. It can be hard to generate demand based on delivery alone, especially without any name recognition or storefront. Many of CloudKitchens’ ghost kitchen peers have disappeared as a result.
But operators also said they got little to no support from CloudKitchens after it sold them on success. They left feeling they’d been misled and at times mistreated.
We can’t help but have similar concerns about Atoms and its lofty plans for the restaurant industry.
We’re willing to be proven wrong, if the company would talk to us or even answer some simple questions. Atoms, after all, is no longer in stealth mode. So what does it still have to hide?