Sidewalk Robots Over Self-Driving Cars

This view says the winning delivery robot in cities will look less like a self driving car and more like a bike courier with software. In dense neighborhoods, the hard part is not cruising down a road, it is getting right up to the restaurant, loading fast, stopping without blocking traffic, and handing off close to the customer. A small robot that can use bike lanes, sidewalks, shoulders, and crosswalks fits that workflow better than a car sized vehicle.

• Coco built around merchant access, not road autonomy alone. Its robots show up like a driver, merchants load them with no training, and the fleet routes across sidewalks, bike lanes, and roads using live maps of blocked paths, traffic lights, flooding, and construction.
  1 sacra
• The closest comparables split by environment. Starship also uses small robots for last mile food and package delivery, while Nuro uses larger road vehicles that top out around 25 mph and were designed as zero occupant cars for road delivery. That makes Coco and Starship a better fit for dense storefront corridors, while Nuro fits more car native routes.
  2 sacra 3 sacra
• Platforms are already treating delivery as a mixed fleet problem. DoorDash works with Coco and Serve, and Uber works with Coco, Serve, Nuro, Cartken, and Avride. That matters because platforms want the cheapest vehicle that can actually complete each order, and in crowded urban cores the small robot has better stopping, parking, and curbside economics than an autonomous car.
  4 sacra

This pushes the market toward city specific robot networks rather than one universal autonomous vehicle. The companies that win will be the ones that can run dense urban routes reliably, keep teleoperation labor low, and plug into merchant workflows without asking cities or restaurants to rebuild infrastructure around them.