There's no question that the idea of an autonomous or "self-driving" car has a great deal of appeal. There's also no question that a world in which most of the vehicles on the roads are automated will be safer and more efficient than today's jammed highways.
But when is that world coming? Some tech enthusiasts would have us believe that a self-driving future is just around the corner. But investors hoping to ride this trend should consider the possibility that it will be many years before the idealized self-driving future will be a reality.
The future vision is compelling
When people talk about the promise autonomous or "self-driving" cars, they generally mean vehicles that both drive themselves and communicate with the other vehicles and infrastructure around them.
Once the roads are flooded with these vehicles, the argument goes, accidents and traffic jams will be greatly reduced -- and travel by car will be safer, swifter, and more pleasant.
That all sounds true to me. Companies (and regulators) are already working hard to bring about that future. But there's a catch: To get all of the great benefits, most of the cars on the road have to have self-driving (and intercommunication) capabilities.
That's why I think that a fully self-driving future is probably still a long way off, even though self driving cars are already heading to market.
Self-driving cars are already emerging ...
It's possible to argue that the first self-driving car has already arrived -- but only for an extremely limited definition of "self-driving."
Daimler's (NASDAQOTH: DDAIF) Mercedes-Benz already has an extremely limited self-driving feature available on a couple of models: It can take the wheel in stop-and-go highway traffic. But it only works up to 37 miles per hour, and it doesn't work when you're not bumper-to-bumper on a clearly marked highway.
General Motors, Tesla Motors (NASDAQ: TSLA), and a few other automakers are expected to release similar systems over the next 18 months or so. In fact, Tesla has promised that a software update for existing Model S sedans will enable some limited self-driving abilities in the near future, possibly before the end of 2015.
For now, these first systems will mostly be gadgets in expensive luxury cars. Think of them as increasingly smart versions of cruise control rather than as robots that can drive your car.
But the expectation is that these systems will get more sophisticated over time, and automakers will gradually add them to mainstream models as costs come down. Meanwhile, the U.S. government is already working on standards for vehicle-to-vehicle and vehicle-to-infrastructure communications, and a few automakers -- again, starting with luxury brands -- are rolling out some very limited capabilities.
... but a self-driving future is still many years away
I've talked to several auto-industry executives about the likely timeline for self-driving cars. All agree that fully self-driving vehicles won't be available for a while yet. That's partly because the government is still figuring out the rules for such vehicles, and partly because the technology still has to overcome some big technical challenges: For instance, rain can be very confusing to a self-driving car's sensors.
These executives say that while many manufacturers have promised self-driving cars by 2020, it's likely to be several years after that before true, fully self-driving vehicles are available to the mass market.
But even if those cars were to hit the market tomorrow, there's another reason it'll take a long time before that utopian self-driving future emerges: It's what the auto industry calls "replacement rate."
Here's the key figure: The average vehicle on U.S. roads today is over 11 years old. Vehicles built today are much more durable and reliable than the vehicles of 20 or 30 years ago. People (and businesses, and governments) are keeping them longer.
It's possible that most every new car on the U.S. market will have self-driving capabilities within a decade. But even if that happens, it's likely that it will take another decade before most of the cars on U.S. roads have that ability.
That is, unless some sort of big disruption happens.
Even the Apple Car won't change that
The alternative view is that the emergence of self-driving technology leads to a sudden move away from the idea of private car ownership. Instead of dealing with the hassles of owning (and driving, and parking, and fixing, and insuring) cars, people will opt to subscribe to an automated car service that can take them wherever they need to go, without the hassles.
I think it's likely that Apple (among others) is looking to create such a service. There has been a lot of speculation about an Apple car, but I don't think Apple wants to enter the auto business. I think the company is exploring the idea of creating a premium car service using automated electric cars.
Uber is also believed to be working on such a service, and it's a safe bet that other companies are as well. But even if that alternative vision is the one that prevails, it'll almost certainly be many years before it spreads beyond the city environments where Uber is succeeding with human drivers now. And in the meantime, people outside those cities will still be buying (and probably driving, at least part of the time) familiar-looking cars and trucks from the established automakers.
Either way, that self-driving future is more than just a few years away.