Waymo has filed a request to the California Public Utilities Commission to expand robotaxi service in the San Francisco Bay Area and Los Angeles region. In the SFBA, it grows from just San Francisco to the whole peninsula, all the way to Sunnyvale but not including Marin, the East Bay and Santa Clara/Cupertino/San Jose. The LA area includes everything north and west of Compton, but not the San Fernando valley.
The SFBA service area includes Google HQ (but for odd reasons, not its NW side.) On a personal note, it includes my home, which is right at the edge of the area. It includes interstate 280 and 101, important now that Waymo is operating on freeways with employees in Phoenix. Operation would be 24/7, in rain or fog, and up to 65mph.
This expansion is huge, and a huge step in demonstrating Waymo’s ability to scale quickly to larger areas, but curiously not quite enough to perform the next really important test Waymo must do, which is learn how it can act as a car replacement. Right now, it’s operating like a taxi/Uber service, and it is not their ambition to just be another Uber (though they work with Uber to do exactly that in Phoenix.)
To go beyond being another Uber, Robotaxi services must convince people to give up ownership of a car (possibly the 2nd or 3rd car in a household) and replace it with use of a combination or robotaxi, robotaxi-enhanced transit, taxi and other modes. That’s where the real money is, and where the world-changing is.
The challenge with this service area is that San Franciscans often travel to Oakland/Berkeley and Marin, and sometimes beyond. Silicon Valley denizens from Palo Alto to Sunnyvale often travel to San Jose, Santa Clara, Cupertino and a few other towns. So there it’s not practical for people in these two core zones to give up a first car for the Waymo service, though they might drop a 2nd or 3rd car, particularly if they can easily and affordably use human-driven services or frictionless car rental to reach more remote locations. (The latter is needed for any trips out of town, and while it could be used for a trip from Mountain View to San Jose, it’s not as pleasant for that.)
The Los Angeles service area might work well for those in downtown, Hollywood, Santa Monica and similar areas, but they will still have travel outside those areas that must be made convenient and affordable. (Regular $2.50/mile Uber prices are not viewed as very affordable unless rare.) In 2-3 car households however, you can arrange so that if somebody wants to travel outside the service area, they get the owned car, and others use Waymo. Then you just need something special on days with 2 parties want to leave the area, which will be rare for many families.
Frictionless car rental can be an answer. For this, small depots are set up along the border of the service area. A rider takes a robotaxi to the parking lot where a rental car is waiting. The transition into the rental car is ideally very smooth — it recognizes the rider’s phone and lets him or her drive off with no paperwork, with an equally simple return, on par with or better than typical automated carshare. As long as the detour is very minor, this can be a great solution for all long distance trips. Indeed, the rental cars, driven by the customer, can be SUVs or pickup trucks for special trips, making the service more useful than ownership of a sedan.
It is not necessary to please everybody, though. In a wide service area you must limit customer at first, using a quota and waiting list, to keep wait times reasonable until such time as you can greatly increase the fleet. As such you can offer the service only to those who very rarely leave the service area. Certainly nobody who works outside the service area would get on the early rider list as they would be unable to commute—unless they want to test the service on people who do all but commute with it.
The service areas include SFO and LAX airports. Waymo’s first effort to get access to SFO airport (which sits on an enclave of land controlled by the city of San Francisco) ran into problems, the battle between San Francisco and the robotaxis is far from over. Indeed, while the Waymo area includes San Francisco, it’s quite possible they will de-emphasize use for residents, though people in the suburbs of course want to travel to SF frequently. It’s easy to let a limited number of non-residents who are flying in make use of the service while they visit—visitors are a great demographic for robataxi use, though of course this is similar to what Uber does.
The limit of 65 will be problematic on interstate 280. Human drivers routinely go 80mph or more on this road, and staying the limit will restrict a car to the right lane, which is also the most chaotic. In the first years, Waymo had a dial in the car which could be used to command the car to exceed the limit (on human authority and responsibility) but that experiment ended. It probably needs to reappear but that’s not going to happen any time soon. Even so, for those riding from the peninsula to SF, even a 65mph ride where you don’t have to drive would win over personally driving at 80mph, unless the customer is running late. Perversely this means that forbidding the vehicle from speeding the way humans do makes the roads more dangerous, as those in a hurry switch to private cars and have more crashes because of the restriction.
Pricing
The big question will be how they price the service. At Uber-like prices of $2.50-$3/mile it’s hard to use it as a car replacement. The longest trip from Sunnyvale to San Francisco could cost $100 to $150 each way, which many would reject. More likely customers would like a flat monthly fee that compares with or beats what they pay for car ownership today, if you add lease payments, insurance, gasoline, maintenance and parking, among others. Or a flat fee that matches the first two, and per mile fees that match the latter elements to exactly mirror (and beat) car ownership. Of course, by eliminating the driving, one can cost a bit more than car ownership (until personally owned robocars become an option that offers this.)
Parking is key, but you tend to only pay for parking in the downtowns of these areas. Once you have to include parking cost, it’s much easier to beat the private car.
Also possible is pairing with transit. While the SFBA’s Caltrain leaves a fair bit to be desired, it would be a lot better if you could use a robotaxi to get you to it and pick you up at your destination train station. (Waymo HQ is located next to a Caltrain station.) Even better, Waymo could run buses such as the Google bus fleet that already does this job for Google and Waymo employees. The buses could provide greatly superior service, even though they don’t get the private right-of-way the train gets (that’s how bad the train is, though shortly it will improve when it goes electric.)
In particular, with buses you can group passengers together so the bus trip is non-stop—the train stops at every station or most stations.
It’s not said but Waymo might very well promote its service to Waymo and Google employees, to supplement the free transportation system Google already provides to these staff. They are ideal early users, though in fact they are clearly not enough as Waymo can and has done this without needing CPUC permission.
“Third car” replacement is more readily possible in these service areas. Third cars are often for teenage children. Even if the robotaxi service is somewhat expensive, many parents (who trust robocars) might find it more acceptable to give their children access to a robotaxi service rather than give them their own car. Particularly if the deal is that the teens can drive (if they have a licence, of course) one of the parental cars when one is declared spare, and use the service otherwise. Or they may prefer the kids don’t drive or get a licence, as teens are the most at-risk drivers on the road. Waymo would have to allow solo minors in the car. At some point it might also allow those from 12-16, who can’t get a licence, to ride on certain trips as well. That’s probably not in this early stage, since the consequences of a crash with a minor could be dire.
It will be interesting to see just where Waymo goes if and when this service area is granted.
