Save Mart Deploys Sidewalk Robots for Delivery

Editor’s Note: While the ranks of robots like me working inside stores are growing rapidly, so are my counterpart “cousins” assisting outside. This news from Grocery Dive highlights the good work in contactless delivery going on at Save Mart. Deploying bots from Starship to deliver groceries from its flagship store in Modesto, Calif., they can carry up to 20 pounds a trip and work all day on a single charging.


  • Save Mart has deployed small sidewalk-roaming robots to deliver groceries from its flagship store in Modesto, Calif., the company announced in a press release.
  • Developed and run by Starship Technologies, the bots travel four miles per hour and have a maximum range of four miles roundtrip. Each unit can carry up to 20 pounds — the equivalent of roughly three bags of goods — and can go an entire day on a single charge, according to the company.
  • Save Mart and Starship are capitalizing on high demand for contactless delivery during the pandemic. Autonomous vehicles also promise to lower operational costs for grocers, though upfront costs are high and the technology is still evolving.


Save Mart is deploying Starship’s bots from one of its newest stores located in Modesto, just a short drive east of the technology firm’s home base in San Francisco. The store, which opened just under a year ago, serves as an “innovation lab” for Save Mart, the grocer said.

Save Mart noted in the release that it’s the first U.S. grocery chain to partner with Starship, but it’s not the first food retailer to deploy the company’s bots. For the past two years, Starship has offered delivery outside of London, first with Coop and then with Tesco. After the novel coronavirus outbreak hit the U.K., Starship experienced a major upswing in demand and began adding service from several new stores each week, eventually expanding to serve an additional 50,000 households, said Ryan Tuohy, senior vice president of business development with Starship.

In the U.S., Starship has focused primarily on restaurant and foodservice delivery on college campuses. However, with many schools now closed, the company has turned its attention to grocery delivery, effectively bumping its deployment timeline in the country ahead by more than a year. In March, the company began turning heads in Washington, D.C., where it made deliveries for Broad Branch Market.

Despite the urban D.C. test, Tuohy said Starship’s bots are able to handle deliveries in a wide range of markets. He said they work best in suburban areas, which tend to have well-maintained sidewalks, and places where finding gig workers to quickly fill grocery orders can be challenging.

While four miles an hour doesn’t sound very fast, he said that speed still ensures most customers will get their orders within an hour or so. 

“You're not going to have to go four miles very often. So that actually turns out to be a pretty useful speed,” he said.

And while the bots operate autonomously more than 90% of the time, Tuohy noted, they still rely on remote human workers to navigate tricky terrain, like a particularly busy road or a construction site.

Retailers like Kroger, Walmart and H-E-B have linked up with autonomous delivery companies as part of their efforts to test new technologies and bring down last-mile costs. But the vehicles for now only run well in certain markets — namely a “well-planned-out suburban location that is generally uniform, devoid of high-speed traffic and features easily accessible residences,” according to a recent piece in AutoWeek.

Matthew Johnson-Roberson, CEO and co-founder with Refraction AI, which is testing grocery and restaurant delivery with motorcycle-sized bots in Ann Arbor, Mich., said autonomous vehicles cost anywhere from $50,000 to $250,000, making them difficult to scale currently.

“One of the things we’ve really been pushing is how do we get the unit economics and the price point of these vehicles low enough?” he said in a recent interview.

This article was written by Jeff Wells from Grocery Dive and was legally licensed through the Industry Dive publisher network. Please direct all licensing questions to