An update on the progress and challenges of autonomous ground-based delivery robots
Estimates of the size of the global market for robotic and autonomous delivery vary widely, but most analysts agree that the category is growing fast from a relatively small base.
MarketsandMarkets projects the global delivery robots market at about $796 million in 2025, reaching $3.24 billion by 2030, representing a compound annual growth rate of approximately 32 percent.
Some analysts peg the segment more conservatively – $400 million in 2025 growing to $770 million by 2029 at a CAGR of around 18 percent – illustrating the uncertainty around scale, scope (food v parcel), and definitions (sidewalk v road-going pods).
For last-mile “robot” studies that bundle multiple form factors, estimates range higher (into the low billions by 2030), but methodologies differ.
All that said, for engineers and investors, the upshot is clear: deployments are still concentrated in specific use cases (campuses, select neighborhoods), but the commercial curve is bending upward as autonomy stacks improve and unit economics inch closer to human courier parity.
Why ground robots? The persistent last-mile problem
Urban logistics remains dominated by short-haul, low-value trips. Robots attack cost (labor, idle time), speed (short batching), and sustainability (small electric platforms).
Compared with vans, sidewalk bots and low-speed road pods are quieter, lighter, and potentially safer – if they can reliably navigate cluttered spaces and win public trust.
The 2025 technology snapshot
- Perception and planning: Multi-camera arrays, radars, and LiDAR feed Level-4-style stacks; remote teleops kick in for edge cases.
- Form factors: 6-wheeled sidewalk bots for food/groceries; larger road-legal pods for parcels; airport and mall variants for controlled environments.
- Operations: Fleets lean on remote assistance, geofencing, and staging hubs; cloud dispatch integrates with restaurant POS and courier platforms.
- Safety and incidents: Low-speed designs minimize harm, but mishaps (for example, a robot bumping a pedestrian) highlight the need for better human-robot interaction and incident response.
Sidewalks v streets: Where they actually run
Most scaling is still under way in constrained geographies – university campuses, business parks, and selected districts, where curb cuts, crossings, and density favor robots.
City-wide coverage remains rare due to infrastructure variances, rules, and vandalism risks. Campus food delivery continues to be the dominant “live” workload in North America and parts of Europe, with pockets of neighborhood service in Los Angeles, Chicago, Miami and beyond.
Rules, liability, and insurance
Regulatory posture is highly local. In the US, low-speed, lightweight designs help, and purpose-built pods have received targeted exemptions (for example, Nuro’s precedent for specific design features).
Insurers are warming to incident-rate data, but sidewalk accessibility (for example, American Disabilities Act compliance) and right-of-way remain flashpoints.
Economics
Unit economics hinge on utilization (orders/robot/day), remote-assist minutes per trip, service radius, and battery swap logistics. Partnerships with platforms (Uber Eats, DoorDash, Grubhub) can compress go-to-market costs but shave margins.
Scaling from pilots to thousands of daily deliveries requires reliable autonomy in messy urban conditions and robust field ops.
Competitive landscape: 12 companies to watch
The companies listed below are ranked according to 2024-2025 deployment momentum, partnerships, and operational signals.
1. Starship Technologies – sidewalk scale leader
The category’s volume front-runner; surpassed 8 million autonomous deliveries in 2025 and reports operations across 100+ service areas.
Where it runs: US, UK, EU; strong campus footprint and growing city zones.
Technology: Mature Level-4-style stack, dense real-world data, robust human-robot interaction playbook (clear signals, polite prompts, and easy access to a remote operator).
2. Serve Robotics – platform partner for Uber Eats
Multi-year agreement to deploy up to 2,000 robots on Uber Eats; fleet of more than 400 by mid-2025, with delivery volume up approximately 78 percent quarter-on-quarter in Q2. Public company signals (supply, operations) are improving.
Business: Manufacturing partnership (Magna) and high-profile restaurant brand pilots boost visibility.
3. Nuro – road-going pods and exemptions
Purpose-built street pods with the first US Department of Transportation exemption for certain design features; now collaborating on next-gen robotaxi/delivery platforms.
Recent development: Broader autonomous vehicles tie-ups suggest converging robotaxi and delivery capabilities; scaling remains difficult.
4. Coco Robotics – dense-city operations (Los Angeles, Chicago)
Claims 1 million miles, 1,000+ robots produced, and expanding with major platforms (DoorDash/Wolt), plus fresh funding to scale thousands of vehicles.
Reality check: Field reports still show edge-case struggles in crowded zones – typical of sidewalk operations.
5. Cartken – campus specialist with platform partnerships
Active with Grubhub on campuses and Uber Eats in city pilots; manufacturing collaboration with Magna. TechCrunch reports a 2025 pivot to more industrial use cases while maintaining delivery ops.
6. Kiwibot (rebranding toward Robot.com) – campus networks at scale
Broad campus presence via Sodexo and financing with Kineo to grow fleet; recent rebrand signals multi-industry ambitions.
7. Ottonomy – controlled-environment delivery (airports, malls)
Level-4-aiming “Ottobots” in airports (for example, Munich T2) and hospitals; strong fit for semi-structured venues.
8. Avride – campus deliveries via Grubhub
Newer entrant deploying robots on US campuses starting with Ohio State; founders trace back to Yandex autonomous vehicle heritage.
9. Refraction AI – mid-sized road robots (status: limited footprint)
Notable for “REV-1” mid-speed road form factor; funding smaller than peers and footprint has fluctuated.
10. Boxbot – automation for parcel staging and robotic last-mile
Focus on automated micro-hubs and sidewalk/curb-adjacent workflows; an ops-centric approach that complements robots on the street.
11. Eliport (Barcelona) – European pilot activity
Ground robots targeting urban logistics; early-stage with ongoing development in EU smart-city contexts.
12. Historical or paused programs
Amazon Scout (paused 2022), FedEx Roxo (ended 2022): high-profile retreats underscored the operational complexity of sidewalk autonomy at scale, but their trials informed today’s safety and ops standards.
Four hard problems still blocking city-wide scale
- Edge-case navigation: Construction zones, narrow sidewalks, grade changes, and occlusions still trigger remote assist too often, capping utilization.
- Weather and durability: Rain, snow, and heat stress sensors, traction, and batteries; weather-hardening adds weight and cost.
- Curb management: Safe, legal crossing and hand-off at the doorstep remain non-trivial without consistent infrastructure.
- Security and vandalism: Theft, tampering, and prank interactions require better locks, alarms, and camera coverage, and thoughtful human-robot interaction.
What’s working right now
- Campuses and semi-closed districts: predictable maps, high order density, cooperative facilities teams.
cartken.com - Platform partnerships: direct integration with Uber Eats, DoorDash/Wolt, and Grubhub compresses merchant onboarding and boosts demand density.
- Operations discipline: rapid field servicing, battery swaps, and 24/7 remote assist centers lift uptime.
The road and sidewalk ahead
Expect steady expansion in campuses, hospitals, airports, and mixed-use districts, plus selective neighborhood coverage where cities support curb policy and micromobility.
The line between delivery pods and robotaxis may blur as autonomous vehicle stacks converge (for example, Nuro’s collaborations), but last-mile economics will still favor small, light, slow platforms for short-hop jobs.
Walk in progress
Last-mile delivery robots are past the novelty phase and well under way toward operational maturity in controlled environments. City-wide autonomy remains a work in progress, but 2024-2025 data show rising fleets, bigger partnerships, and clearer playbooks.
For engineers, the frontier is reducing remote-assist minutes and hardening hardware; for investors, the signal to watch is not press-release counts but deliveries per robot per day and route completion without intervention.
