Prediction: 95% of 2030 Travel by Self-Driving Cars

“By 2030,” says a new report from a group that calls itself RethinkX, “95% of U.S. passenger miles traveled will be served by on-demand autonomous electric vehicles owned by fleets, not individuals.” The Antiplanner is more optimistic about the rapid growth of self-driving cars than most, but RethinkX’s prediction is more dramatic than anything the Antiplanner has said.

As recognized in this more moderate report from UC Davis, RethinkX’s statement is really three predictions in one: first, about self-driving cars; second, about what powers those cars; and third, about who owns those cars. I think 95 percent by 2030 is optimistic for any one of these predictions, much less all of them.

First, the decision about what powers cars is completely, 100 percent independent of the decision about whether humans or computers drive cars. So long as the United States gets most of its electricity from fossil fuels, even natural gas, the environmental benefits from converting to electric cars is negligible, especially since we can make gasoline-powered cars more fuel-efficient.

To the extent that we can convert electrical power to renewable sources of energy, that energy should be dedicated to existing users of electricity, not brand-new users. So there is no moral imperative to shift to electric vehicles, and I suspect the advantages of gasoline over electricity in terms of range and costs will still exist in 2030. So I strongly doubt that 95 percent of our vehicles will be electric powered by 2030.

RethinkX argues that electric vehicles will lead to the collapse of the oil industry, so there is no need to allow companies to do fracking on public lands or build the Keystone or Dakota Access pipelines. I am dubious about that.

Second, the decision about owned vs. shared cars is partly dependent on whether humans or computers drive cars, but that is not the only factor. If computer drive the cars, car-sharing services will be much less expensive than using a taxi, Uber, or Lyft today. But that doesn’t mean everyone will give up car ownership.

Some people live in such remote areas that shared car services aren’t practical. The Census Bureau says that 20 percent of Americans live in rural areas or communities of fewer than 2,500 people. Owning a self-driving car will actually enable more people to live in such remote areas. So that 20-percent-plus will prefer ownership over sharing. In larger communities, many people will prefer ownership simply because they like their personal space.

There is also an economic condition to consider. Car owners only have to pay the marginal cost of driving each additional trip. Car sharers have to pay the average cost of driving every trip. Anyone who thinks they will drive more than the average number of miles will be tempted to own rather than share. For all these reasons reasons, I suspect only about half of vehicle use will be shared.

RethinkX projects that electric cars will last for 500,000 to a million miles or more. That means the capital cost will be a much smaller share of the total cost compared with today’s cars that only last around 200,000 miles. They project that the average cost of sharing a car will then be less than the marginal cost of owning a car, so everyone will share. This doesn’t address people who live in remote areas or who prefer their personal space, and it doesn’t really address the average vs. marginal cost issue because, if shared electric vehicles can last more than 500,000 miles, so can owned electric vehicles. So I’m not persuaded.

Finally, as much as I look forward to self-driving cars dominating the roads, the notion that 95 percent of travel will be by self-driving cars in 2030 is even more optimistic than I can believe. For one thing, 10 percent of our travel is by air, and I don’t think we are going to give up much of that to self-driving cars.

A second factor is that self-driving cars will be dependent on extremely precise maps. On one hand, such maps will allow them to travel on snowy roads where stripes and shoulders are obscured. On the other hand, this means they can only travel on roads that have been precisely mapped.

A company called Here that is making such maps is co-owned by Mercedes, Volkswagen, BMW, and Intel. Ford has invested in a similar company called Civil Maps. Waymo and possibly Apple are making their own maps.

The United States has 2.7 million miles of paved roads and 1.3 million miles of unpaved roads. Maps would also have to cover every driveway and parking area. Despite competition between the various companies, I doubt that all these roads be mapped by 2030, so anyone using other roads will need a human-driven car. Even when all the roads are mapped, some people who want to go off road will need human-driven cars.

The Antiplanner is more optimistic than most about the rate that autonomous cars will penetrate the market because many recent cars can be retrofitted to become self-driving. But most cars made before 2010 or so, and many cars made since then, can’t be retrofitted. For example, I doubt that any cars with manual transmissions can be retrofitted. Even as late as 2030, there will be lots of cars on the road that aren’t self-driving. People aren’t going to toss them all in the junkyard in order to fulfill RethinkX’s prediction.

When the Antiplanner started writing about self-driving cars in 2009, many people thought I was nuts, so I am gratified that so many other people have jumped on the bandwagon. But my major point has been that the effects of self-driving cars are unpredictable. This means we shouldn’t spend public dollars on technologies that self-driving cars will render obsolete. But it doesn’t mean government is omniscient enough to overrule private investments.

RethinkX’s predictions are interesting, but no one should base long-term policy on them. Instead, public officials and government agencies should concentrate on solving today’s problems today, and let the future take care of itself.

Share

14 thoughts on “Prediction: 95% of 2030 Travel by Self-Driving Cars

  1. JimKarlock

    Saying that electric cars last 500,000 miles is leaving out a few expensive things that still wear out and cost lots of $$$.
    Front end rebuild – 100-200k miles??, $1500
    Rugs, seats, windshield pitting, general interior, rust problems, wheel bearings, door hinges, paint job, power windows, and anything else that moves!

    Basically what you save with electric are brake jobs and drive train rebuilds.
    thanks
    JK

  2. OFP2003

    Well, if a fleet-owned self-driving car is as dirty as the WMATA trains – you can forget it!!

    Question, how much of the self-driving tech is dependent on GPS? and good GPS signals? Hopefully engineers are building in a “work around” for times the GPS network is down (solar flare, etc).

  3. jon

    When I had three kids and was on a five year self-funded sabbatical I was using a 4-door Toyota Echo. Over the five years that we owned that car, if I remember my calculations right, I spent on average $50/month including depreciation, insurance, and repairs but not gas. I could put my car seats for the kids in the car and keep them in there. I could go camping “off road” or in more rural areas.

    Yeah, with those kind of numbers I don’t see why people wouldn’t want to own their own car. I think it might be ideal for single people or people without kids but now that I have 4 kids, small cars are not an option. I paid $9000 for my current vehicle. I doubt that it would be cost effective with renting.

    Also, I wouldn’t want a pure electric car until they can go 1000 miles and can fill up in 10 minutes or less. Since we are and always have been a single car family the vehicle we own must be versatile. Sometimes we take trips with just a single person in it, since it is more cost effective to do that than to buy a second vehicle. It would be nice to go down to Phoenix in the van ride share but the costs are so high that it is more cost effect to just drive my SUV not to mention the convenience of being able to drive around myself when I’m down there.

    Had one other point but it has slipped my mind. 🙂 I’m excited about self-driving cars, but I just don’t see how they would work without the option of taking over occasionally when you want to go off the beaten path.

  4. msetty

    Once a lot of people figure out what the various private mapping companies want to do, there will be huge resistance. Already, in Germany and Austria, you can’t get the street view in Google Maps because the Germans felt pictures of everyone’s house was an invasion of privacy.

    In this country, I suspect most homeowners, ranchers and other property owners will not allow the mapping at the detail needed for fully self-driving cars. And forget about detailed mapping of “gated” “communities” (sic).

    I suppose you could fly drones over a lot of these places and generate 3-D maps, but translating such maps into those usable for self-driving cars would be yet another complex algorithm on top of many others, opening up even more possibilities for failure and accidents. Ultimately for most people, not worth the trouble or extra cost.

  5. CapitalistRoader

    Rethinking Transportation 2020-2030
    The Disruption of Transportation and the Collapse of the Internal-Combustion Vehicle and Oil Industries

    It sounds like they’ve got lots off good info in the article. So, why did they have to saddle it with such a ridiculous title? Natural gas and nuclear will be generating the majority of electricity in 2030 and probably for the remainder of the century. Will electric cars slowly gain market share? Sure, as batteries get better along with more charging infrastructure. But the change won’t happen that fast. I’m guessing the natural gas hybrids will come to dominate car and truck propulsion in 2030, if only because we’ve got so much gas and the efficiency difference between electric cars (burn gas to make electricity to send it over miles of wire to charge the car battery ) and hybrid natgas IC cars (burn gas to charge the car battery) just isn’t that great.

    This is heartening though:

    “TaaS Pool will be cheaper and more convenient than most forms of public transportation. This will not only blur the distinction between public and private transportation but will also most likely lead to a virtual merger between them. We expect that TaaS vehicles will be largely differentiated by size, with two-, four- or eight-seaters and up to 20- or even 40-seaters in the TaaS Pool market. ”

    No more expensive trains! I do think though that they’re overlooking peoples’ need for privacy during their commutes, which is the only time they’re truly alone now. I expect shared cars will have individual compartments with privacy glass and adjustable HVAC controls.

  6. msetty

    Hey, CapitalistRoader, “…peoples’ need for privacy…” will trump (sic) a lot of what is needed for self-driving vehicles such as super-detailed maps, as I explained.

    So you don’t mind that the “TaaS” companies will compile even more data on your every move and activity, on top of what they already do with smartphones and other devices on the Internet? You’re so concerned about privacy??! Ha!

    “TaaS Pool will be cheaper and more convenient than most forms of public transportation. This will not only blur the distinction between public and private transportation but will also most likely lead to a virtual merger between them. We expect that TaaS vehicles will be largely differentiated by size, with two-, four- or eight-seaters and up to 20- or even 40-seaters in the TaaS Pool market. ”

    Here is the ultimate delusion of the futurists and techno-philiacs: that somehow, the time-consuming deviations of demand-responsive ride-sharing services–whether driven by robots or humans–that is required to provide door-to-door services to individuals, will go away with automation.

    Beyond a few people, say up to a half-dozen, it will still be more economic to run fixed route transit services–whether driven by robots or humans. An automated bus running in a straight line–say, over a fixed, well-defined 7-mile route–would take typically 20 minutes in the usual U.S. suburban environment of 2,000-4,000 per square mile (21 mph). If the bus costs $1.50 per mile to operate (electricity, maintenance, amortization of capital costs, bus stops, etc.) you’d only need 2-3 riders to keep costs below $5.00 per ride. If the same vehicle made 2-3 deviations over the route, you’d add 5-6 minutes per deviation, making it a 35-40 minute trip rather than 20 minutes straight-line.

    Sure, for each person picked up at their curb, they save ~5 minutes walk time to a bus stop, but you inconvenience the other folks on the vehicle, making a significantly longer trip, even with access time to a bus stop.

    In denser environments with volumes of more than 4,000/square mile, passenger volumes will be higher and deviations even more onerous.

    Put another way, technology can’t defeat geometry, as Jarrett Walker points out, even if some of his commenters are even more idiot savants than the typical futurist. http://humantransit.org/2017/04/the-virtues-of-impossibility.html.

  7. Barleyman

    Rush Hour anyone?

    Everyone is travelling at the same time to and from different locations. The poor, overcrowded or if you like socially-and-environmentally responsible already go by bus or train or subway or bicycle or foot (or some combination of the above), but everyone currently driving to work wants to leave and arrive at different locations at more or less the same time. You just can’t do that with fewer vehicles.

    I could possibly believe that in 13 years I’ll be replacing my human driven vehicle with an electric self-driving (at least some of the time) vehicle, if the costs come down and efficiency of electric vehicles goes up, but I don’t want to have to wait 1/2 an hour for a fleet vehicle to fetch me from home and drop me off at work every day, and again in reverse each evening.

    Car ownership is not so financially burdensome in the USA that people will automatically give up their current convenience and habit just to save money. I can ride the bus for free now but I don’t, because it would take at least 1 1/2 hours each way to commute, not counting time walking to and waiting at the bus stop, against a 20 minute drive. Yes, I pay the costs of car ownership, and an annual parking fee, but the bus option is not a desirable lifestyle choice.

  8. The Antiplanner Post author

    OFP2003,

    Driverless cars will rely very little on GPS. If you tell a car to go from where you are to some other destination, it might use GPS to map a route. From then on, it will rely on internal maps. If GPS networks go down, it probably would not cause much of a glitch to most driverless operations.

  9. CapitalistRoader

    M. Setty: …peoples’ need for privacy…” will trump (sic) a lot of what is needed for self-driving vehicles such as super-detailed maps, as I explained.

    I just overlooked your other post because most of it just doesn’t make any sense to me, but if you insist:

    Once a lot of people figure out what the various private mapping companies want to do, there will be huge resistance. Already, in Germany and Austria, you can’t get the street view in Google Maps because the Germans felt pictures of everyone’s house was an invasion of privacy…In this country, I suspect most homeowners, ranchers and other property owners will not allow the mapping at the detail needed for fully self-driving cars. And forget about detailed mapping of “gated” “communities” (sic).

    There is no expectation of privacy from a public road. In the US, anyone can take a picture of anything from a public road, well, except one can’t purposely try to take a picture of the interior of someone’s bedroom i.e., where one as an expectation of privacy. But cars won’t be driving into people’s bedrooms, they’ll be driving on roads.

    Do you have a problem with Google Maps now? I don’t. Currently the front of my houses are displayed with decent resolution. How can that harm/how does that reduce my privacy? Anyone can walk down my sidewalk now a take a picture of my house. It’s not against the law nor do I think it will ever be against the law.

    In the US, anyway. Citizens in EU countries are much less free than US citizens. Most European countries have blasphemy laws, for God’s sake, including Germany and Austria. They really are quite backwards when it comes to individual freedom. And why would anyone spend time gathering location data from AVs when they already collect the same data from cell phones? Does riding in an AV somehow make location data more valuable or unique?

    Regarding the economics of mass collective transit vs. private or shared autonomous cars, you mentioned the main reason the demand for AV’s will probably eliminate collective transit for all but the densest places: Fixed routes. That is, you have to go to the transportation vs. an AV coming to you. It’s a much pleasant way to conduct your life on a daily basis.

    Lastly, the geometry argument. Existing highways absolutely will have more carrying capacity with robot servos vs. us meat servos. I can easily imagine having twenty feet–one car length–between AV’s traveling 80mph vs. typically at least two car lengths now and probably closer to four. Doesn’t that automatically double a highway’s carrying capacity? Roadway carrying capactiy of dense, urban roadways will obviously see much less improvement but the whole point is that fewer people will need to be packed into urban areas since 1) the cost of automobile transit drops and 2) opportunity cost of automobile transit time is reduced because you will be able to do other, more valuable things with the time instead of piloting the car.

  10. prk166


    Nearly 100 million existing
    vehicles will be abandoned as they become economically unviable.
    ” ~Rethinkx

    Wow. This happens all the time. I used to drive an old car and carried the title with me. If it broke down on the road, I was calling a junkyard to come take it. I’d give them in title in exchange for the taking it off my hands.

    The problem with this claim, like too many others, of Rethinkx is that they give very few details behind it. 100 million sounds like a lot. 100 million over a decade would be a slower rate than currently occurs in the US. To mean what I suspect they’re talking about, people would in mass walk away from perfectly functioning automobiles that they’ve paid for and whose incremental trip costs are near nill, at least until something on them breaks. This is a highly questionable claim, to say the least.

  11. prk166


    In this country, I suspect most homeowners, ranchers and other property owners will not allow the mapping at the detail needed for fully self-driving cars. And forget about detailed mapping of “gated” “communities” (sic).
    ” ~msetty

    I’m not sure what you’re getting at here. This has already occurred with 3D mapping of buildings based on those 2D images, automatic sign recognition, road condition assessed via AI, et al.

  12. CapitalistRoader

    Hybrids more energy efficient than EVs:

    The below “40 mpg, EPA combined” table shows, high-mileage E10 vehicles, including hybrids, such as the 52 mpg Toyota Prius, have greater energy efficiency than EVs, and only slightly greater CO2 emissions than EVs. It would be much less costly and quicker to significantly increase the US hybrid fleet, than to build out the EV fleet, which is still in its infancy, and would require major, expensive changes to supporting infrastructures.
    Comparison of Energy Efficiency and CO2 of Gasoline and Electric Vehicles, The Energy Collective, 11 May 2017

  13. prk166


    Once a lot of people figure out what the various private mapping companies want to do, there will be huge resistance. Already, in Germany and Austria, you can’t get the street view in Google Maps because the Germans felt pictures of everyone’s house was an invasion of privacy.
    ” ~msetty
    msetty, they are there and gathering data. They just do not publish it via streetview. Most of the data is still available for other uses. The real key to these pictures is not what can be seen, but the 3D mapping that can be auto generated from it’s data.

    Furthermore, other sources are able to share pictures similar to streetview.

    https://www.mapillary.com/app/stats/85633111?lat=50.926034785989344&lng=9.927274294672713&z=5.652212169864692


    In this country, I suspect most homeowners, ranchers and other property owners will not allow the mapping at the detail needed for fully self-driving cars. And forget about detailed mapping of “gated” “communities” (sic).
    ” ~msetty

    This gathering of data has been occurring for decades. The key is public, not private property. The law and precedence for this is pretty clear at this point.

    Keep in mind that these driverless cars are gathering the data they need as they go. If they don’t already have the chicken, they’ll make their own egg.

Leave a Reply