Some Say Self-Driving Robotaxi Isn’t A Business; Billions Are Being Bet That It Is

Most of the biggest names in self-driving cars are aiming to make money selling Robotaxi service — most quickly described as a self-driving Uber

UBER
-style service where you can summon a car with an app on your phone and ride elsewhere for a reasonable fee, possibly combined with “sharing” in some form, such as the style of UberPool or forms of on-demand transit.

This is the plan of Waymo, Cruise, Amazon

AMZN
/Zoox, Ford/Argo and many others. It was obviously the plan of Uber ATG before it sold to Aurora, and Lyft

LYFT
L5 before it sold to Toyota. It’s even a big part of the plan of Tesla

TSLA
. While Tesla intends to sell consumer cars, it promotes that they will become part of the “Tesla Network” where customer cars temporarily hire out to give rides, and also that Tesla will buy 3-year old cars as they come off lease to put into a robotaxi fleet.

Indeed, last week, Dan Ammann, CEO of GM/Cruise predicted that they would see $50 billion in sales from 1 million vehicles, largely from robotaxi, by 2030. (Current GM revenues are $122B.)

Almost all the top players like robotaxi plans because if you sell the ride, you sell the whole automotive value chain — vehicle, fuel, maintenance, insurance and more — and control the customer and all the revenue for this whole chain. You become the automobile industry, the oil industry, the car repair industry, the auto insurance industry, the auto loan industry, the car cleaning industry, the parking industry and more. You decide where the money goes. It’s the catbird seat. There are suppliers but they jump to your call.

Some players want to continue as car companies, selling cars to consumers as they have always done, but with the new luxury — and eventually standard — feature of self-driving. That’s harder, because a car you sell has to go everywhere you want to sell it, while a taxi can be quite profitable working a modest sized service area. The taxi comes home every night and you can watch over it and improve it. The private car is out of your hands. A private car that can just do highways and main roads is nice, particularly for the commuter or long drives, but it’s not the value of the car that can get to most addresses.

At the closing (virtual) session of the annual Transportation Research Board self-driving conference (the oldest one in the industry) 3 reporters, including Forbes’ own transportation editor Alan Ohnsman, debated this topic. The surprising common view was that robotaxi did not look like a good business, certainly not the grand business that all those who have invested tens of billions into it believe.

You can also read this analysis from the Financial Times on how opinions have shifted between Robotaxi and Driver Assist business plans. This includes statements from ADAS-approach converts like Austin

Tens of billions is nothing compared to the theoretical value. Around the world people are spending close to $5 Trillion on personal ground transport. Ownership of even a small piece of that pie, in time, is easily worth almost any investment today. Waymo has deployed a pilot project in suburban Phoenix which actually operates and charges Uber-like fees to members of the public, but it’s hardly a place to test this business. Almost everybody there has private cars, it’s about as suburby as it can get.

In September, Waymo started service in San Francisco, but excluded the lucrative downtown area where most Uber rides take place. That’s still far from ideal, though they surely plan to add it as soon as they can.

So how do robotaxi players capture the immense value they hope for?

Not in the country

It’s unlikely robotaxi service will arrive in rural locations for a long time. Aside from the fact that car ownership is a must in these places, the density is low enough that it’s harder to provide a robotaxi to pick you up in just a few minutes. Many rural addresses have a driveway that’s quite long, and the effort to make sure the car can drive one driveway may be hard to justify for just one customer.

The suburbs are not as bad. The density is high enough and the roads are faster than in the downtowns, so a ride can come to you on short notice. You don’t have to do more with the driveways than pause in them.

It will be a long time though before companies attempt to deploy enough vehicles in a rural area to provide quick response. While cars can travel more quickly on rural roads, so you don’t need nearly the density, it will take effort to get anybody a car in 5 minutes. You need something like that to hit the real business goal, which is…

Car replacement

If you robotaxi service is just a taxi service, it can be a good taxi service, but it still remains nothing more than that. Internet taxis like Uber and Lyft do far more business than yellow taxis ever did, so it’s not a small business, but it’s not the big competitor for traditional car ownership you want to be. Robotaxis will probably quickly capture all tourist travel (including rental cars for use in the city) and many other things.

To do that, you need to convince people to give up ownership of a car and use your service a lot. That’s a tall order, but it’s made less tall by the fact that they don’t have to give up all their cars — early business will come from people giving up 2nd and 3rd cars — and they don’t have to use your service exclusively, there just has to be a range of services which meet the needs the car used to meet.

Replacing a 3rd car is pretty easy. In fact a lot of 3rd cars are there for the kids, or for specialty needs (like when you need a truck.) Even a 2nd car can be fairly easy. A robotaxi service will offer a variety of special vehicles. It’s a truck this morning, a minivan at lunch, and SUV on the weekend and a mini-car on the commute home, something that an owned car can’t do.

To replace a car, other services are needed, which may include transit and traditional car rental for trips out of town. It may also include human driven taxi service for trips outside the robotaxi service area. As long as these unusual trips are rare, they need not add much cost or trouble. If a customer wants to commute every day outside the robotaxi service area, they probably aren’t a great customer.

To assist with this, robotaxi companies will experiment with different pricing models. While taxis tend to price by the mile, and robotaxis will surely offer that (at a lower price) many customers will like subscription models which more easily compare with car ownership. Many car owners today pay a monthly lease/loan fee, annual fees for maintenance and insurance, and a weekly fill-up of fuel. It’s possible to price just like that, or through other variations — and deliver a superior price and better service. The price can even include reasonable usage of rental cars and human-driven taxis to assure the vehicle completely replaces the lost car.

Some customers won’t be so keen on complete replacement. There are already many people who sit on the line between car ownership and alternatives like transit, ebikes, carshare and taxi. Low cost robotaxi service can easily push them over the line. You need to push enough over the line to have a real world-changing business.

And while transit people fear this, robotaxis will take some business from transit. They hate it because transit is designed to cost the same whether it runs full or empty, and anything that loses a passenger is bad news. Future transit (which will use self-driving vehicles too) will be able to be more flexible on vehicle size so that this equation is no longer true. Vehicle size will respond to demand and remain efficient.

Waymo has faced criticism that Chandler, AZ is just not enough of a challenge any more — and they have started a service in non-downtown San Francisco. Other pilots underway in China and elsewhere are all more challenging. It’s also true that Chandler just isn’t a great candidate for car replacement. The best places for car replacement will naturally be places that people already give up a car for things like transit, e-Bikes or Uber. There you just have to nudge them over an edge that others have already been nudged over. It’s also handy if parking is hard to find, not just at destinations but in homes. Many urban locations have either no garage or a one-car garage. In some places one must pay for a parking space (though building codes of the past actually forced apartment towers to provide lots of parking, today they are starting to reverse that, and forbid lots of parking.)

How to make a profit

Certainly nobody is close to profit on robotaxi service. In fact, critics point out that firms like Uber and its ilk bleed money, though there are suggestions Uber might report its first “profit” next quarter, at least on an EBITDA basis, and that Didi and Lyft might also be on the way after recovering from the hard hit of the Pandemic which cut taxi service to the bone. Uber charges around $2/mile and keeps roughly 50 cents of that from the drivers (though it pays a fair chunk of that back as driver incentives.) Still, for that 50 cents/mile it has few per-mile costs — mainly insurance and moving a few bits around. 50 cents a mile is a common all-in cost for using a privately owned car, not counting parking.

To win the business, the robotaxi service has a few options:

1. It can be better, even though more expensive. Obviously not having to drive and thus being able to get things done is a huge plus, as is being dropped at the door of your destination and not having to find parking.
2. It has some other ways to be better, including offering a variety of vehicles to meet the need, and even letting you have 2 cars at once as long as you need that rarely. The ability to pick up friends coming to visit you. Things owning a single car can’t do.
3. It can be cheaper. At first, this probably only happens in places where parking costs money, or for people who are below average in car use. While that average private car might cost 50 cents/mile, if use is light, each mile can be more. Freeing up garage space can also be important, offering storage or even finishing the garage as a room in some places.
4. It can eventually be cheaper by using low cost single-person vehicles for single person trips — which are about 80% of trips. It also continues to be better while getting cheaper.

Once it gets to being both better and cheaper, it has the potential to take a lot of business. Most of the 2nd cars, and a fair number of the 1st cars. Of course, there are some it won’t ever take — those who don’t like the technology, those who like to keep a lot of stuff in a private car, and those who feel that car ownership is a big part of their identity. As noted, rural people will be less likely to have service good enough to replace their car, though it’s still possible at the higher price needed to justify a large enough fleet to give short waits.

How much profit?

It remains puzzling that Uber is struggling to make a profit when it takes 50 cents/mile from the typical ride with low per-ride costs. 5 years ago, I released this spreadsheet on robotaxi economics estimating the per-mile Cost-of-goods-sold (COGS) and possible retail prices of double that COGS. Uber has not come close to that sort of gross margin, even when you remove the Uber driver’s portion of the fee from the equation. Clearly at this stage the cost of running a big company like Uber with its technology is being allowed (though the use of investor’s money) to get too high. The stock market seems happy, nonetheless.

Some items in the cost remain uncertain. One is the cost of regularly cleaning vehicles. Uber style vehicles have a human driver in them who cleans them on the spot, and who also prevents the passenger from feeling the temptation to get too messy. A robotaxi has no human watching, and when it needs cleaning it needs to waste time and resources heading to a car cleaning depot. The cost for the cleaning can and will be billed to the passenger who messed up the vehicle, but will that be enough deterrent? Car-share faces the same challenge and mostly deals with it. Classic car rental includes cleaning at the depot, and it’s included in the price, but many people return the cars very much in need of a cleaning.

As is typical, software and hardware R&D are not included in the COGS. Neither are things like staff and management and marketing and more. In theory their cost does not go up with the number of miles of travel. That’s not entirely true. New territory with QA, mapping and tweaks does add new costs, and you need to increase territory to increase miles. Robocar software will never be “done” and so while you can’t attribute it to any mile of travel, its ongoing rather than a capital investment like the physical car whose depreciation is part of the COGS.

The ability to have 80% of one’s trips in a small vehicle which can retail rides for under 40 cents/mile still predicts the potential for a profitable service — cheaper than car ownership for all but those who get the heaviest use from their cars, and without the burdens of parking (at home or at the destination) and the return of free time without the driving task. This offering, both better and cheaper for most, is why we see the billions of dollars being invested, chasing trillions of revenue. Even at $1/mile (a likely price after early maturity) the offering is superior and only modestly more expensive than ownership for the mainstream, and should attract many.

This doesn’t consider the other large segment of the robotaxi world, namely shared robotaxis and vans able to carry 3 to 15 people on group rides, moving groups and replacing much of public transit. Their economics are different but also interesting. By meeting public goals, they may also see subsidies which can rewrite any set of economics too.

These pilots will help answer those questions when they get large enough to be seen as car replacement, and are allowed to operate as full commercial services. Some answers are still to be learned, but the prospects continue to be grand from a financial standpoint — even before one considers the options that come from improved safety and enabling new styles of moving in our cities.