No, cars won’t be driving us in the near future. Here’s why not

In May, GM spent $1 billion to buy Cruise Automation, a small startup with promising self-driving software. Last year, Qualcomm’s bought NXP Semiconductors for a cool $39 billion, in large part because Qualcomm wanted to buy its technology for making chips for autonomous vehicles. And in March, Intel paid $15 billion for an Israeli company, Mobileye, which makes sensors for driverless cars.

Some of the world’s richest, most innovative, and coolest companies—like Apple, Baidu, Google, IntelTesla, and Uber—are pouring money into autonomous vehicles (AVs) and related technologies.  Waymo, Google’s project, has driverless cars on the road in four US locations, and they have collectively gone more than 3 million miles.  Intel sees a $7 trillion future for AVs by 2050.

AVs have two big advantages. While most people think they’re great drivers, many are not. In the United States alone, almost 100 people a day are killed on the roads, and 94 percent of the crashes are due to human error. AVs show reliably better judgment than humans, and they never drive while texting or drinking. Second, a road full of AVs can accommodate more vehicles, because they can travel closer together, lessening the need for new roads and speeding up traffic. The latter is no small thing, considering US motorists spend more than 6.9 billion hours a year stuck.

No wonder, then, that people look at these trends and see a future—a very near future—of highways buzzing with super-computer-enabled AVs where people are entirely dispensable (see image). A Morgan Stanley analyst in 2014, for example, saw fully-fledged driverless cars—no human intervention needed—on the road by 2019 and 100 percent AVs in as little as 20 years. Another consultancy, ABI Research, was slightly less bullish in its 2013 forecast, but sees half of sales in the form of AVs by 2032. Ford said last year that it could have an AV in mass production by 2021. NuTonomy, an independent manufacturer, says it will have AVs taxis in 10 cities by 2020; Audi, Ford, Nissan, Toyota, and Volkswagen all see AVs on the market by the end of the decade. Former Uber CEO says the company’s entire fleet will be autonomous by 2030.

So, that seems pretty persuasive. But I am not so sure. The enthusiasts, investors, and experts—and these groups overlap considerably—may be exactly right. But I am also struck by how similar what I am hearing now about AVs is to the way people were talking about electric vehicles (EVs) not that long ago.

In late 2010, for example, one expert prediction was that by 2013, 200,000 electric cars would be sold in the US, and in 2015, 280,000; in fact, the figures were 96,600 and 119,000, respectively. Carlos Ghosn, the well-respected CEO of Nissan, said in 2011, that by 2016, there would be1.6 million Renault-Nissan EVs on the road; that forecast was off by more than 80 percent. President Obama saw a million EVs on American roads by 2016; the real figure is fewer than 300,000.

I believe that EVs and AVs will be a big part of the future of the car. Just not so fast. For EVs, the big issues have been cost and technology. Both are improving, as batteries get better and cheaper, but are not there yet. When Georgia’s $5,000 tax credit for EVs expired, for example, sales immediately slumped; ditto for sales in Denmark and Hong Kongwhen incentives expired. These are signs that people will not pay more upfront for an EV even if the total costs of ownership over time are competitive. “Range anxiety”—the fear of running out of juice and no way to refill—is still a factor. New models, such as the 2017 Chevy Bolt, will go more than 200 miles on a single charge, which should go a long way to help. Ultimately, though, as Ghosn has said, the infrastructure needs to catch up with the hope (and the hype) to build consumer confidence.

For AVs, the challenges are similar in some ways, and very different in others. The differences are that AVs will require new regulatory and legal guidelines, and new forms of logistics. In the litigious United States, figuring out liability in car crashes—and these will, of course, still happen—is going to be a doozy. There are thorny ethical issues, too. Here is just one hideous hypothetical: What should an AV do in a situation in which the choice is to hurt a bystander, the occupants of the car, or another car? What kind of principles should be imbedded into the AV’s algorithms?

Moreover, AVs require sophisticated mapping and data analytics to do their thing. Not to belabor the obvious, but the United States is a really big country. As the CEO of Shift, a used car sales service, put it, “private sector innovators will need to work with legislators and their staffs to secure permission to survey roads and impede on daily lives of the citizenry.” And while mapping, say, North America, will be challenging enough, the obstacles will be much greater in other parts of the world.

Finally, as cool and as positive as AVs could be, there are always downsides, and if there is mobilization against them, things could slow down sharply. For example, there are about 3.5 million truck drivers in the United States, and they regard AVs with great suspicion, for obvious reasons. Finn Murphy, a long-haul trucker and author of The Long Haul: A Trucker’s Tales of Life on the Road, notes that the topic of AVs is cropping up in truck stops all over the country.

The similarities, however, could be even more important and of these, the most important is that when it comes to their wheels, Americans kind of like them the way they are. They’re used to the internal combustion engine, and not that eager to change. That was the message at a recent Energy Information Administration conference. The results of a national survey last year by Kelley Blue Book found that—perhaps unsurprisingly—American consumers are not nearly as excited about the car of the future as the experts are.

For EVs, the survey saw growth, but still less than 4 percent of the market by 2023 (or about where it was supposed to have been in 2013). Even among EV owners, the enthusiasm is underwhelming; only 53 percent said they would buy one again (compared to 82 percent for traditional cars); the figure was 31 percent for plug-in electric owners. With prices at the gas pump in the United States still comparatively low, one of the major reasons to go hybrid or electric is less relevant. Considering that the Chevy Bolt is priced at almost $40,000 and Tesla’s forthcoming Model 3 around $35,000—about $10,000 more than a comparable conventional car—there is a ways to go. According to the EIA’s own forecasts, in 2040, gasoline-only vehicles will still account for the vast majority of light-duty passenger car and truck sales.

People would look more closely at EVs if the sticker price came down—52 percent said so. And here’s something that is genuinely surprising: For all the coverage EVs have been getting, when asked to name one, the first two models that the respondents named—the Prius and the Volt—are not, in fact, EVs. My McKinsey colleagues anticipate that uptake of EVs will be faster in richer, denser cities, with strict emissions standards and generous incentives. That sounds right—and that also sounds like a limited market.

As for AVs, there does appear to be genuine interest, particularly among young people–as long, that is, they can keep their hands on the wheel when they want to. And considering that one of the great advantages of AVs is supposed to be safety, ironically 74 percent of Americans surveyed said that they didn’t think fully autonomous cars were safe. In all cases, moreover, they don’t want to pay for it. In one survey, the amount that consumers said they would be willing to pay for advanced technologies has declined by a third since 2014, and a significant minority doesn’t want to pay at all.

We have just passed the 10th anniversary mark for the iPhone—a great example of how fast habits can change when a product delivers what consumers want—even if they don’t know they want it at first. But the iPhone is also a great example of how badly even very smart people can get it wrong when it comes to predicting the future: Clayton Christenson and Steve Ballmer, for two, thought it was no big deal.

But unlike the smartphone, which was an entirely new concept, consumers know what it is like to have a car. So it’s worth considering what they seem to be saying—that they are intrigued but skeptical about AVs, and that EVs are still too limited and too expensive. This is what they are telling surveys—and certainly what they are saying with their wallets.

McKinsey research sees fully autonomous cars being 15 percent of sales by 2030, “once technological and regulatory issues have been resolved.” That might be spot on. My point is simply that those technological and regulatory issues are substantial—and so are those related to consumer attitudes and acceptance.

Yes, there is a future for both EVs and AVs. Volvo’s recent announcement that it is going to transition toward an all-EV platform beginning in 2019 is just one more indication of this. But as the recent history of EVs shows, just because a cool technology exists doesn’t mean it is going to take over the market at speed and scale. The same is likely to be true of AVs: This is going to be long road, and a bumpy one.