In 2040 self-driving cars will still be way outnumbered by proper cars

autonomous cars
A showroom demo in Berlin of the drivereless Volkswagen Sedric concept car, recently. Image credit: Sergey Kohl /

ITEM: A new forecast from IHS Markit expects sales of autonomous vehicles to pass 33 million a year by 2040 – but most of those will be in China, the US and Europe. And that will only comprise just over a quarter of total new car sales.

The big driver for autonomous vehicles in the early days will be ride-hailing or mobility-as-a-service, not individuals buying new cars, says Egil Juliussen, Ph.D. and director of automotive technology research at IHS Markit:

“The first autonomous vehicle volumes — beyond retrofit test vehicles — will arrive in 2019 through driverless mobility services […] Volumes will surpass 51,000 units in 2021 when personally owned autonomous cars reach individual buyers for the first time, and IHS Markit forecasts estimate nearly 1 million units will be sold in 2025 across shared fleets and individually owned cars.”

The US will be the first market to start selling autonomous vehicles, with total volumes forecast to reach 7.4 million units per year in 2040. However, China is expected to close that gap quickly once regulations for autonomous vehicle testing and deployment are in place, after which the industry will hit 14.5 million autonomous vehicle sales in 2040.

Most of those sales will come from mobility service fleets, not personal ownership. In Europe, IHS Markit expects the opposite phenomenon due to Europe’s strict regulations regarding ride-hailing. Consequently, of the forecasted 5.5 million autonomous vehicle sales annually in 2040 in Europe, the majority will be for consumers rather than mobility services.

Which means that of the 33 million autonomous vehicles sold annually in 2040, over 27.4 million will be sold in the US, China and Europe. That leaves just nearly 6.3 million per year for the rest of the world – most of them likely in markets like Japan, South Korea, Singapore, Australia and Canada, where conditions for autonomous driving are more favorable in terms of factors like regulations, the presence and strength of the local auto industry, and the complexity of local driving conditions – although some or all of those could change over the next 23 years.

These forecasted numbers raise a few points regarding all the excitement over self-driving vehicles.

For a start, you can forget about those PowerPoint panaceas about orderly grids of autonomous cars safely ferrying everyone to and from the office or wherever. According to IHS Markit, autonomous vehicles will account for 26% of new car sales in 2040 – which means the other 74% will be standard human-driven vehicles, or at least vehicles with limited autonomous abilities (like self parallel parking or next-gen cruise control).

Factor that in with the older cars that are likely to still be in use, and you have a scenario where even by 2040, the number of self-driving cars on the road will be dwarfed by the number of human-driven cars. Given all the hype about how self-driving cars are safer than human-driven ones, that number doesn’t exactly inspire optimism. And that’s assuming autonomous cars are as safe as advertised (I’m betting they won’t be, if only because I’m a firm believer in Murphy’s Law).

There may be other factors in play, such as the current assumption that as ride-sharing (automated or otherwise) becomes more common and reliable, fewer people will buy cars. But according to a recent report from research house Bernstein and Business insider, the total number of cars on the road worldwide will almost double from 1.1 billion in 2015 to 2 billion by 2040.

Another thing becoming clear is that the future of autonomous vehicles will be nowhere close to evenly distributed. Even in the more advanced countries, autonomous cars will be in the minority – in emerging markets, they’ll be an urban rarity.

Granted, a lot can happen between now and 2040 that will influence how the self-driving vehicle saga plays out. But if IHS Markit’s numbers are anywhere close to right – and IHS Markit is one of the more optimistic analyst firms touting the disruptive opportunities that autonomous vehicles and mobility-as-a-service will create – for most of the world the future of self-driving cars is a lot farther away than next week’s Lyft demos at CES will probably make it look.


  1. I am still looking forward to seeing the first autonomous motor cycle and wondering how all those bikes are going to stay out of the way of autonomous cars!!
    Autonomous cars are one thing. Them succeeding without taking ALL the old manual stuff (and pedestrians, especially Jaywalkers) off the road is still inconceivable for me.
    Do we have autonomous legal judges for accidents too?

  2. I feel this is rather conservative prediction. As we all know, new things often come slowly first, but then there is a turning point and things start to happen very rapidly. When I was a kid, I remember a prediction that in 20 years there would be a computer in most of homes. I felt it then very unbelievable. Or how internet and mobile came to everyone, when necessary components were available (of course tech component had existed then a long time). And if we think the history of different new innovations, it is not so clear new things come mainly to richer and more developed countries first. And this is not only about development of cars, but this is very much about development of business and transportation models and life style too. Of course, this is just my personal thinking, without any research or data analytics for this.

    But I’m ready to make a bet, I claim there are more self-driving cars in traffic in 2040 than old fashion cars!

    • Fair points. But I maintain that cars are too expensive and last too long for self-driving cars to see the kind of accelerated uptake that computers and smartphones eventually did.

      That said, the balance might be tipped if the majority of people buying new cars are fleet owners as consumers decide ride-sharing is better than owning a car – they’re likely to buy new cars sooner (like rental car agencies, for example, who typically keep new cars no longer than two years before selling them and buying new ones).

      Still, I’ll take that bet. I’ll be 75 in 2040, so I’ll probably need the money. 🙂

  3. I find it hard to imagine many consumers wanting to own autonomous vehicles. Far more likely that they’ll be ‘auto-taxis’ that will be used on demand – effectively a taxi without a driver. Not too much of a stretch for many of today’s city-dwellers.The majority of people with good reason to buy a car – people in rural or remote locations, say – will still want conventional self-drive cars, partly because the tech infrastructure probably won’t stretch far enough to be trusted, and partly – very importantly in my opinion – because people like driving cars.

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.