Cruise and Waymo, the clear #2 and #1 (respectively) in the Robotaxi race, have recently expanded their service areas for public access rides and driving with nobody in the vehicle. It’s a continued positive milestone in a year that has seen many setbacks for self-driving projects.

1. Cruise announced it will begin service in Chandler Arizona (with some overlap with Waymo’s first service) tonight with service from 7pm to 2 am.
2. Cruise will also start service in a small area around downtown Austin and UT Austin from 10 pm to 5:30 am, the same hours they use in San Francisco.
3. Cruise in November expanded service in San Francisco to all hours and most of the city, except the busiest part of the CBD — but only for use by Cruise employees.
4. Waymo recently expanded service in San Francisco to add the full downtown only for employees and guests. Waymo continues to operate 24 hours/day, but adds safety drivers during certain weather conditions. Waymo service outside the NE quadrant of the city has been available to members of the public.
5. Waymo doubled its service area around downtown Phoenix. Previously it also opened service from there to Phoenix airport, via a stop at the PHX Skytrain people mover that serves all terminals.
6. Waymo has said opening service in Los Angeles is imminent. Waymo is anticipating it will soon get a permit to charge money for rides. The current service, while technically open to the public, has a waiting list.
7. Waymo disclosed it now has 700 robotaxis in operation or testing in is various test territories.

I recently rode in Waymo’s service in the Noe Valley area of San Francisco and the ride was smooth and impressive. There will be more about that ride in a forthcoming article.

Once Waymo allows members of the public access to the central business district, and can charge money, they will have the first commercially viable robotaxi service. (It is possible one of the services in China has reached this milestone but this is more difficult to evaluate.) While they have been open to the public in Chandler for several years now, Chandler just isn’t the sort of city you would build such a business in. Phoenix is so sprawled that it’s still a challenge to have a commercial service with just the downtown.

This is not to say that Waymo’s service will be profitable. Indeed, they still have massive R&D and logistical costs that make that far away. The full San Francisco service area is both a viable taxi market — after all this is where Sidecar, Lyft and Uber were born — but more importantly a potential market for car replacement. A service established here, with lower prices than Uber, could convince San Franciscans to give up ownership of a car. At present, it would probably be a drop from 2 or 3 cars, not a drop to having no cars, but this is a start.

Such a service would not be pure robotaxi. It would also probably need to include:

1. Service to SFO airport (though it can take you to BART which serves the airport.)
2. Easy carshare use, where they take you to a rental or carshare vehicle with frictionless transfer, at the North, East, and two South freeway exits from the service area, and get you a good price.
3. Access to transit, as well as bike share and scooter share
4. Ideally transfer service to any other Waymo service areas, such as the one anticipated for Silicon Valley
5. Occasional use of human-driven Uber-style service at a good price
6. Short wait times

People would give up car ownership for this — they already give up car ownership in cities like San Francisco just through the use of transit/bikeshare/scooters/Uber/Lyft without the convenience and lower cost of a robotaxi. But a service like this might be attractive to a lot more people, to create a viable market.

Cruise is not there yet. Their services are only operating at night for the public, and the difference in difficulty is, to put it both literally and metaphorically, “night and day.” Service with employees is allowed to have many more rough edges, because, unlike the public, employees are not posting videos and online stories of problems — and we are indeed seeing that for the public services.

Both companies are way ahead of other players, who are only operating a few robotaxi services with safety drivers ready behind the wheel. Some companies in China are operating with no safety drivers there. Tesla’s prototype self-driving system is not capable of operation without extremely frequent supervising driver intervention on any roads, anywhere. Zoox is not operating a service. Motional has been operating service with safety drivers in some locations, including in partnership with Uber and Lyft. MobilEye, another leading player, has not yet opened live service.

These expansions illuminate one of the key debate questions in the field, particularly in consideration of Tesla’s approach. All leading robotaxi teams develop and maintain detailed maps of their driving territory, and don’t deploy in a territory (certainly with members of the public) until they have tested and certified their vehicle’s safe operation in that territory. While Tesla does not have self-driving, they are allowing customers to do driver-supervised driving on a large fraction of U.S. roads. The hope is that if they can make self-driving work, it would instantly work everywhere, while the other approach is built one region at a time.

While Tesla wants to drive everywhere, today it can’t drive anywhere. (I know, I own one and have studied many videos.) The other players can actually perform the robotaxi task (which is to be able to operate for tens of thousands of hours without any safety mistakes) but in a limited set of areas.

The question is, which is harder — being able to reach the very high robotaxi safety performance level at all, or expanding the maps and service area? Waymo and Cruise have bet that making it work is the hard part, and scaling maps and certification is the easier part, while Tesla is working as though scaling maps will be too hard and expensive compared to making the driving system work.

The ability of these companies to quickly add territories provides information to help answer this question. Cruise announced it would open in the two new cities by the end of the year, and it pulled it off — though only at night. It seems probable, however, that once the core case of a viable service can be made, we should see a “land rush” as the companies which have viable systems work to claim and expand to new territory.