9% Of Cars To Be Driverless Cars By 2035 – Forbes
So says the consultancy IHS IHS Automotive and I have to say that I think that’s a gross underestimate: or possibly a gross overestimate. Yes, I know, but I do think that it will either be a thoroughly commonplace product by then or we’ll still be struggling to make it work.
Here’s what IHS Automotive says:
Sales of vehicles able to drive themselves will account for 9pc of global auto sales in about two decades, according to a forecast published by auto industry consultant IHS Automotive.
The study focused on autonomous cars, which can drive with “no attention needed by the driver”, IHS analyst Egil Juliussen said. Such cars are not currently available for sale, but IHS predicts they will be available in around 2025.
IHS expects global sales of self-driving cars in 2025 to be 230,000 – less than 1pc of the 115m cars expected to be sold that year.
But by 2035, sales of self-driving cars will reach 11.8m, or 9pc of the 129m global auto sales expected that year, said Mr Juliussen.
I’m afraid that I just don’t see it. For there are, to my mind, really only two possibilities here. The first is that for either technological or legal (and that second is to my mind much the greater barrier) reasons we’ll just not be able to make driverless cars work as we want them to. In which case adoption will be a bit like that of solar cells in the 1970s. A few purchasers out of some sort of pioneering spirit but the rest of the world simply ignoring the technology.
Alternatively, we do manage to crack the technology and that legal stuff. And as I say, I think that second, about tort law, responsibility for crashes and so on, is going to be the real deal breaker. But we crack it all: so why won’t near all cars then carry the technology?
We might, perhaps, think that the package that makes a car driverless will be very expensive but that ignores two things. The first being Moore’s Law: we can make driverless cars right now with today’s technologies. Given the speed with which computing power becomes cheaper in 20 years’ time the cost of something computery that we can do now will be spit. I might be exaggerating a little here but a 1994 computer (say, a Pentium III?) cost $1,000 or so in that time’s money. Today we wouldn’t even use a processing chip of that power in a low end feature phone that we’re going to give away when you buy $10 worth of airtime. The second thing is about the costs of mass manufacturing. There really does come a point in the installation of some gadget in something mass manufactured that it is more expensive to not include it than it is to include it. Simply because the costs of making two different things are greater than the costs of putting whatever it is into everything.
Now I don’t insist that this will happen but a reasonable guess would be that, assuming the whole technology can be made to work, that the marginal cost of installing it in each and every vehicle manufactured will be a couple of hundred dollars. Don’t forget, anything in patent now will be out of patent by 2035. And at that point I just cannot see that the market penetration will be only 9%.
My own prediction is that it will be under 1%, because it’s just not working, or more like 90% because it so obviously is.
Given that 2035 is my approximate checkout time on the threescore and ten count here’s hoping I live to see whether this prediction makes a fool of me or not.