Waymo announced today they would, as promised, expand their service in the suburbs of Phoenix into the downtown area, accepting applicants for an early rider program. Back in March, they opened up service in San Francisco after having done testing there for a long period. Cruise, the self-driving project of GM, also expanded their service area in San Francisco last week to cover much more of the town — though not the downtown and only at night. Cruise’s applications for a permit to charge for rides is in process and expected to be granted by June.
In China though, the action is even faster. In order to impress us, AutoX released this week a drone video of their depot where they have over 1,000 robotaxis serving areas of Shenzhen, including no-human-aboard service in about 1/4 of their 1000 square kilometer territory.
And WeRide released their own video of a side project building robotic street sweepers, but it didn’t just make one or two for a pilot as we might expect from even well capitalized US companies. In Chinese “go big” style they made a whole fleet.
Back in the USA, MobilEye (a unit of Intel
All of this is to say — it’s on. While these are still pilot projects, we are starting to see the first glimpses of the “land rush” that is to come in the 2020s, where players will start to try to claim territory and work to create real working businesses. We aren’t there yet — none of these projects are anywhere close to what would be making money, but they are charging money to work out the details of what that will be like, and putting stakes in the ground of places to operate.
While we see competition in cities like San Francisco, Shenzhen and Beijing, which are the home turfs of these companies, most of this land rush won’t be that competitive, as players move to claim virgin territory rather than go head to head with competitors. Deploying a robotaxi fleet is very capital intensive, and while these are some of the best capitalized companies ever, it’s large even for them.
Companies will start mostly in snow-free areas, though they will soon feel the call of New York City, the world’s largest taxi market. While their goal is not to be simply a better or cheaper replacement for taxis and TNCs, it’s a good place to start — if you can make it anywhere, you can make it there, to reverse the song lyric. Teams will be looking for cities where they can get good business, but also where they can move to the car replacement business where the real money is.
In Chandler, Waymo could create a shuttle service between its two service areas which might allow it to be car replacement for residents of Chandler, who may find they mostly go to downtown and their own area, though most people want occasional trips to the airport and certain other major attractions. Phoenix is so sprawled it’s actually a hard city to do car replacement in, since a car replacement has to have a service area that covers a large fraction of the trips a customer does in their car. That’s much easier in cities where there are already many people who don’t own cars — you just have to push more of them over the threshold by offering great robotaxi service.
The streetsweepers are interesting because it’s one of those side markets which can be made commercial fairly quickly. Some of the earliest commercial deployments came in places like warehouse yards and mines, or on the sidewalks with delivery robots. These solved direct business problems, and could take the simple approach of just stopping if they encounter a problem. Streetsweeping is something that needs to be done regularly, but the vehicles do not drive in traffic. They deliberately go much slower, and can stop and slow go around obstructions when they need to. Since you can run such vehicles 24/7 the savings from pulling out drivers are much more clear, and everybody appreciates cleaner streets. While the vehicles are probably expensive today (and drivers for them not expensive in China) this is something that could be sold to almost any city, particularly in places where drivers are costly.
The vehicle is quite large, and does not yet take advantage of the ability to make very different vehicle shapes when a driver is not involved. That will no doubt come in the future.