What if the Autonomous Car Industry Is Wrong?
Never bet against human nature.
Do you spend time in Silicon Valley or Detroit? If you don’t, know that most conversations involve these talking points: Autonomous cars are inevitable. Almost here. Will be ubiquitous. Save lives. Reduce traffic. Cut pollution. Also, mobility. And sharing. And no one will own cars. FYI, Don't buy Tesla. Tesla sucks. Wait for our stuff.
“Never assume,” my father always said, so let’s follow his advice, deconstruct the clickbait underlying much of the autonomous driving narrative, and ask the question:
What if the autonomous car industry is wrong?
First we have to answer this: What is the autonomous car industry? On one side, we have the universe of Silicon Valley companies trying to figure out how to monetize an immature technology. On the other, we have the universe of legacy car companies terrified the upstarts are going to leave them behind. Since no one knows when or how it will be possible to monetize autonomous cars, they’re investing billions in anything with the words Autonomy or Mobility, catchphrases of a seemingly inevitable future they don't understand.
Are autonomous cars inevitable? Of course. A self-driving car that works on the streets of Mountain View in decent weather? Google has them now. A self-driving car that is 100% guaranteed never to make a mistake, anywhere, in any condition? Not in my lifetime, and I’m not that old.
The billion-dollar question is when will autonomous cars be commercially available? Tesla says 2018. The legacy manufacturers say within five years. The hurdle won’t be the technology, but legal. Where they will be available is another story. College campuses and golf courses don’t count. In order for massive investments to pay off, autonomous cars need to be where the customers are, and that means city centers. How long will it take to convince a patchwork of thousands of cities to allow (if not mandate) them? Longer than five years, which is why Uber bought Otto, whose self-driving trucks we are likely to see adopted long before self-driving cars take over Manhattan.
Will self-driving cars become ubiquitous? That becomes obvious once you’ve digested the most important lesson in Fight Club: On a long enough timeline, the survival rate for everyone drops to zero. Apply this to cars: There are 274,000,000 cars in the United States, give or take, with a turnover of around 17 million annually. If 100% of cars sold today were self-driving, it would take 16 years to get to 100% ubiquity.
Chris Gerdes, Chief Innovation Officer for the DOT, told The
Drive that he thinks 35% of the cars on the road in the United States will be self-driving in ten years.
To get there, 100% of the cars sold would have to be self-driving by 2021.
Not a chance.
Are we approaching a self-driving tipping point? Maybe, but not a global one, or even a national one. When autonomous driving comes to America, it will come in regional pockets determined as much by culture as anything else. Manhattan and sections of California will likely be swamped with autonomous cars. But Texas? Ford and GM could give cars away for free to these proud drivers and never see ubiquity. Montana and much of the West? Expect self-driving semis (e.g. Otto) platooning to turn I-80 into a defacto railroad, with semi-autonomous pickups driven by humans who won’t use the features unless forced, or tired.
Or at all.
Will self-driving cars reduce traffic? Bad news for the self-driving industry on this one. The answer is yes, but only once ubiquity is reached where human driving is banned. Anywhere humans and self-driving cars mix, traffic is going to get much worse, and stay that way. Self-driving cars won’t break the law and will be inherently cautious, which means the presence of even one in a mixed autonomous/human driving environment will reduce speeds to a crawl.
Will self-driving cars save lives? Of course. The state of American driver education actually is deplorable. The debate over autonomous safety is unconscionable, and a waste of time. No one will sell a self-driving car until they are demonstrably safer, and no one will get into one until they are, or appear to be.
Safety, of course, is a moving target.
Tesla has already proven that people will use semi-autonomous technologies like Autopilot before any governmental body has verified their presumed superiority over humans. Does anyone really believe Tesla (or anyone) will sell self-driving technology that’s less safe than the average driver? Of course not. All self-driving cars need to do is appear marginally better than a human driver, and people will get in.
What about mobility? Mobility is what transportation companies call strategies that haven’t been figured out yet. You sell cars, or you have a platform. Specialization works. Mobility? I throw the press releases away.
What about sharing? I’ve never met a Maven or ReachNow customer, but people love Turo, the horizontal sharing platform that owns no cars, but has an amazing fleet of sports cars in California. I love Turo, too, and car rental companies should be scared. Horizontal sharing seems to make sense, and Turo CEO Andre Haddad has written entertainingly on the topic.
Vertical car sharing—where manufacturers own everything from the factory to the insurance policy, infrastructure and the single brand of car waiting to be shared—seems like the worst of all worlds, for everyone. I’ve enjoyed using Daimler’s Car2Go in Brooklyn and Austin, but the recent closure of their San Diego office suggests that all the money in the world can’t make this work.
Will people still own cars? A lot of money has been lost betting against human nature. Cars aren’t merely transportation. For many, cars are transformation. What and how we drive are two axes on the chart of personal expression. As long as people can afford cars, people will own cars. Maybe not as many, but a lot, for a long time. In modern America, cars are the second most important method of self-actualization after homes. For many, the car is more important, because you can take it with you. For many, car ownership comes before home ownership, if the latter ever happens.
When and where it makes sense to use bikes, subways, trains, yellow cabs, Uber, Turo or Hertz, people will do so. But—as soon as they can afford to—people will lease, finance or buy one.
Don’t bet against human nature.
There are fourteen major car companies in the world. No one believes they can all survive, and Morgan Stanley believes only five or six will. The big ones are hedged against any delay in the adoption of self-driving cars. FCA? People have said they were doomed since before the autonomy conversation started. I think it’s still true, but FCA isn’t alone.
The self-driving opportunity is vast, but it’s not a winner take all game. To win, you’ve got to have something no one else has. When autonomy is a Google license away and a few boxe from Delphi, you’re just another player, and it’s back to square one of building cars people want to drive or ride in.
Anyone whose brand is welded to the idea of human driving is protected, which is why Porsche—who have already said they would offer no more than semi-autonomous features—was wise to double down on their brand.
Everyone else who doesn’t already have an autonomy play—or enough cash to wait for the commoditization of autonomy—is toast. Their only bet is licensing Google’s (or Apple’s, once it arrives) platform and becoming the Foxconn of cars. By that logic, the legacy manufacturers investing billions are probably wasting their time. If self-driving cars become legal and Google’s platform is demonstrably safer than Daimler’s, then people will get into anything powered by Google’s.
In which case the old guard has a problem, because there’s only room for so many Foxconns, and traditional manufacturing will become a race to the bottom.
As for Silicon Valley, I’ve heard rumors Japanese execs are in Israel buying any seed stage startup they can. Why? Because they’re terrified of Tesla. I would be, too. If the Model 3 is a success, Tesla’s stock value will be sufficient to make big moves the OEMs won’t like, and the $TSLA shorts are toast.
Google? They can afford to wait for the OEMs to blow it, and find their Foxconn. Apple needs to boogie, but still has time. Uber, Lyft and anyone else losing money on ride-hailing have the most to fear. They have to get the drivers out of those cars, or else.
There is a huge opportunity left in Silicon Valley, but it’s not self-driving cars. It’s aftermarket safety and semi-autonomous technology. Even if Gerdes is right, 65% of cars on the road will still be human driven in ten years. Everyone in the 65% will have seen or ridden in a self-driving car. Every one of them will want to be as safe as possible while keeping their steering wheels, if not their hands on the wheels.
Where are the aftermarket Autopilot-like systems going to come from? I’m sure George Hotz isn’t the only one thinking this way.