Key Takeaways

  • Coco Robotics runs approximately 1,000 sidewalk delivery robots, with plans to produce 20,000 next-generation Coco 2 units for 2027.
  • Battery swapping doubles hardware battery capital expenditure; Coco replaces mobile swap trucks with distributed merchant charging at restaurant sites.
  • Rash views city sidewalks as harder to solve than road driving because sidewalks lack traffic signals, lane markers, and clear legal conventions.
  • Coco signed a multi-year, bi-directionally exclusive agreement with DoorDash and Deliveroo to fuel its UK national rollout.

The CapEx Math Behind Battery Swapping

Most micro-mobility and hardware fleet operators treat battery swapping as standard operations. Electric scooters, mopeds, and small delivery drones often run on swappable packs. A field technician drives a van, swaps depleted packs for fresh cells, and returns the empties to a central hub.

Rash rejects that model for sidewalk bots. “Ideally, you don't want to be going around swapping batteries because it means you're doubling at least doubling your battery capex,” Rash explains.

If your fleet requires hot-swappable batteries, you cannot buy one pack per vehicle. You must buy at least two: one running inside the chassis on the street, and one charging on a rack in a depot. Add in transit buffers, replacements, and degraded cells, and your battery procurement budget inflates immediately. You also take on the ongoing payroll, fuel, and vehicle depreciation of a mobile swapping operation.

Coco avoids that trap through distributed merchant charging. Instead of running swap vans or central warehouses, Coco docks the robots directly at partner storefronts. The vehicle charges while waiting for food prep or between lunch and dinner delivery rushes. This keeps the battery count strictly at one pack per robot and eliminates dedicated swap labor from the operating budget.

Sidewalks Defeat Standard Driving Models

Road autonomy companies poured billions into solving highways and street grids. Those systems rely on well-defined boundary conditions: painted asphalt lanes, curbs, speed limits, and sequenced traffic lights. If a car behaves abnormally, human drivers generally follow predictable patterns or stop.

Sidewalks offer no such baseline. Rash states: “The interesting thing about city sidewalks is they're the most complex, dynamic, not rulebased environment in the world.”

A sidewalk changes continuously without notice. Children dart unpredictably, dogs cross on thin leashes, construction signs block access ramps, and pedestrians walk against foot traffic while staring at phones. There are no lane markings, speed limits, or right-of-way rules governing pedestrian sidewalks. A delivery robot cannot rely on neat lane graphs or standard intersection logic. It must negotiate informal human social norms in real time, making navigational benchmarking far more volatile than tracking a vehicle down an eight-lane highway.

Scaling to 20,000 Units with Aggregator Partnerships

Hardware startups often run out of money by trying to build demand and delivery infrastructure simultaneously. Setting up an autonomous fleet requires capital. Building a consumer food delivery marketplace at the same time burns through cash twice as fast.

Coco avoided the consumer marketplace trap. The company operates about 1,000 robots, primarily in Los Angeles, but is preparing to manufacture 20,000 Coco 2 robots for 2027. To fill that fleet with actual delivery volume, Rash partnered with the platforms that already own restaurant relationships and consumer orders.

Coco announced an agreement with DoorDash and Deliveroo to expand across the UK. Rash described it as “a multi-year bi-directionally exclusive partnership to bring Cocos all throughout the UK and bring it to some real scale.”

The partnership removes customer acquisition risk. Coco does not need an app on every consumer phone in London or Manchester. Deliveroo and DoorDash route existing short-distance restaurant deliveries directly to Coco 2 units. Coco supplies the hardware and operational routing; the aggregators supply the order volume.

What to Do With This

Audit your physical fleet unit economics before committing to field operations. If you run hardware assets that require ongoing maintenance or recharging, calculate the capital required to purchase duplicate parts versus installing charging infrastructure directly at customer pickup points. Run the numbers on a single-battery architecture with merchant-hosted plugs this week. If distributed customer charging drops your equipment budget by 40 percent, kill the mobile maintenance van before you hire the driver.