Waymo Pushes Into San Francisco

Waymo announced today they are ramping up their efforts in San Francisco. No ride service yet, like in Chandler, AZ, but a clear indication that one is planned soon. This is no big secret — Waymo opened a large depot in the Dogpatch area of town and its vehicles have been seen cruising those streets for some time. Waymo talks about how their sensors (which are custom made by Waymo) have been tuned and positioned for the problems of the urban environment, including people randomly crossing the street.

What’s significant about this push is that since Waymo has been the acknowledged #1 in the space since their launch, GM’s Cruise has always touted that by testing almost exclusively in San Francisco, they were going after the real problem. It’s certainly likely that the first place for robotaxi service to bloom is in dense cities, and it’s the easiest place to get people to switch from car ownership to exclusive use of robotaxis combined with transit and other modes. Indeed, people in cities have already traded in car ownership for use of ride-hail like Uber

UBER


UBER


UBER


UBER


UBER


UBER
and Lyf

LYFT
t

LYFT


LYFT


LYFT


LYFT


LYFT
. That’s less common in the suburbs and rare in rural areas.

Waymo took the baby-steps approach, deploying first in one of the easiest places in the world. Cruise tried a harder city first, and now the race is on to see who will first deploy in San Francisco. Cruise only recently demonstrated operating a vehicle with no safety driver behind the wheel, but their demonstration was modest — done late at night in a quiet part of town not much harder to drive than the Arizona streets where Waymo now has a production service open to the public. In addition, they had an operator in the passenger seat. Waymo has as yet shown no operation without safety drivers in California.

San Francisco, with more complex streets, lots of pedestrians and transit vehicles as well as hills and occasional fog, is the hardest city in California to tackle. It’s still probably easier than most eastern cities such as Boston, and has no snow. And the more chaotic and older cities of the rest of the world offer greater challenges.

Waymo began (back when I worked there) mostly in Silicon Valley, around Google HQ. That’s pretty easy territory. Many teams have begun with something even easier, namely freeways, and these are the focus of all trucking startups. Many teams without huge resources have begun in simpler environments, such as private communities (Voyage) or campuses. In December, Starship revealed it had already done one million paid autonomous deliveries by sticking to sidewalks and campuses. Zoox and Aptiv

APTV


APTV


APTV


APTV
/Motional have gone after Las Vegas, which has the wide and simple streets of the west but a huge taxi business.

Competition City

San Francisco looks primed to be the first city with actual robotaxi competition. Generally, as robotaxi firms start going into production, they will be able to deploy in virgin territory, with no competition. Why go head to head against somebody else where there’s an equally good city where you can be alone. Competition means lowering prices and working harder. But some cities are too plum to give up, and San Francisco will be one, simply because Silicon Valley is the capital of robocars. A few others, like London and Shanghai may see competition earlier — and the other pearl of the taxi universe, New York. Robotaxis are not really competition for taxis, they plan to be much more, but a city that already runs on taxis is a good place to start.

Rich companies may also elect to compete simply to learn from it. They will lose money not to cede prime territory. They need to be cheaper than Uber at $2/mile but need not match private car driving at 50 cents/mile if they don’t have competition. I suspect we’ll see a target around $1/mile to be half the price of an Uber, but less where there is competition. For the first year, San Francisco is the place one may want to be.

One other company — Tesla — dreams it will be able to operate a “Tesla Network” where you can summon a robotaxi which is either Tesla-owned or is a private owner’s care hired out to the network. This depends on Tesla’s claim it will soon have cars that can operate with nobody in them on almost any street. Most people in the industry view this as an extreme claim. If they do it, such a service could spread almost as fast as Uber did, to ten thousand cities in under 10 years. People aren’t holding their breath on that, but if they can do it, they can justify that crazy valuation.