DoorDash and Serve Robotics announced a multi-year partnership in October 2025. Eligible Los Angeles orders from participating merchants could be assigned to Serve’s sidewalk robots, with other markets planned. DoorDash’s delivery platform chooses among human Dashers, robots, drones, and other methods according to the order and location.
Serve had completed more than 100,000 deliveries across its network from more than 2,500 restaurants, according to the companies’ announcement. That is Serve’s cumulative history across customers and cities, not DoorDash order volume from this partnership. It proves that robots can complete deliveries in some operating areas; it does not establish profitability or citywide coverage.
The economic question is narrower than the announcement suggests. A sidewalk robot can be useful on a short trip with a suitable route, an accessible handoff, and an order that fits. It still needs fleet staff, charging, repairs, remote help, and enough nearby orders to avoid long empty trips. The partners did not publish cost per completed DoorDash delivery, utilization, or the share of orders eligible for a robot.
DoorDash is also developing Dot and testing Waymo vehicle delivery in Phoenix. Those modes have different ranges, road rules, payloads, and handoff problems. The value of DoorDash’s dispatch layer depends on choosing the right mode for actual orders, not on counting how many autonomous partners it has signed.
The next useful disclosures would be completed DoorDash orders by mode, service areas, failure and intervention rates, and total cost per delivery. Without them, claims that robots cost a fraction of human delivery or will take a fixed share of urban orders are forecasts, not observed results.