Methodology
How the rankings work
Every AVFox ranking answers the same question: how much autonomous service does a company run today? This page explains the yardstick they share. Each ranking's own page explains who gets credit for a trip and how its numbers are estimated.
One yardstick
Every score sits on the same scale, so a 4.0 in ADS Providers and a 4.0 in Road Robots mean the same volume: about 2.5 million trips or deliveries a year. A 10 means 10 billion a year, about the number of rides Uber completes worldwide. Uber does not report rides separately from deliveries; its whole platform ran at about 15 billion trips a year at the end of 2025, and we estimate rides at roughly two-thirds of that.
| Score | Trips or deliveries a year |
|---|---|
| 3.0 | Up to 50,000 (commercial launch) |
| 4.0 | 2.5 million |
| 5.0 | 25 million |
| 6.0 | 200 million |
| 7.0 | 1 billion |
| 8.0 | 2.5 billion |
| 9.0 | 5 billion |
| 10.0 | 10 billion |
Equal effort per point
Each point on the scale is meant to take about the same effort to climb. Going from 2.5 million to 25 million trips a year (4 to 5) should be about as hard as going from 2.5 billion to 5 billion (8 to 9), or from 5 billion to 10 billion (9 to 10). Early on, a company can multiply its volume by adding vehicles and cities; at the scale of Uber, doubling is the hard part. So the volume behind each point grows fast at the bottom of the scale and slowly at the top.
A straight line would not work: on a scale from zero to 10 billion, every company ranked today would score below 0.1. Between two anchors, scores move in proportion to the multiple of volume, not the raw amount: Waymo's estimated 30 million trips a year score 5.1 in ADS Providers, and Starship's 2 million deliveries score 3.9 in Road Robots.
Below 3: getting to commercial service
A company scores below 3 until it runs driverless commercial service. Points measure how close it is: approvals, public riders, no one aboard, operating hours and area. Each ranking sets its own points, because launching a robotaxi app takes different steps than launching a sidewalk robot. A score below 3 says nothing about volume.
One trip, credited once per ranking
Each line follows the same trip from different seats. In Mobility, a ride counts once for the app that booked it (Demand), once for the system that drove it (ADS Providers) and once for the company that ran the car (Fleet Operators). In Delivery, an order counts once for the app that took it (Demand) and once for the company whose robot or drone delivered it (Road Robots, Drones). The same company can appear in several rankings with different scores.
Estimates
Few companies publish trips by app, system or operator, so most scores are estimates. We use the best public evidence available, in this order: reported volume for the last 12 months, a recent run rate, then fleet size times typical use. In the Mobility rankings, a figure more than a quarter old is scaled by fleet growth since it was reported. We don't adjust scores for judgment: when the evidence changes, the score changes. Every score shows its calculation and sources under "How we got here".
Rounding, ties and China
Scores show one decimal. At most two companies share a score; a third is listed 0.1 lower, in order of the unrounded score. The exception is 3.0, the launch level, where any number can tie. Demand and Road Robots come in two lists: Excluding China, the default, and Including China. ADS Providers leaves out companies that operate only in China, and Fleet Operators covers rides outside mainland China.
Each ranking's rules
Mobility: Demand · ADS Providers · Fleet Operators
Delivery: Road Robots · Demand · Drones