The Driverless-Car Pile-Up and the Folly of ‘Industrial Policy’
A fleet of Uber self-driving cars at a technology demonstration in September 2016. (Aaron Josefczyk/Reuters)
If our best and brightest venture capitalists can’t pick winners, the government doesn’t stand a chance.
The idea that the government might successfully support and steer innovation is making a comeback as wonks both left and right show a renewed interest in “industrial policy.” But faceless functionaries steering anything from D.C. should terrify us all. Even the most credible, savvy venture capitalists and entrepreneurs fail at an astonishing rate. Why would a bureaucrat with a ton of money do better?
To see how difficult it is to push the frontier, take the coming wave of innovation in the auto industry.
Over the past three years now, I have watched from my perch at the corner of Broadway and Front Street in San Francisco as a small fleet of SUVs suffers the most dreadful punishment outside my office window. Circling and circling, sometimes farther and sometimes closer, but always coming back like two-ton boomerangs, these SUVs have taken the same routes around the same city blocks, every day, day after day.
It would make Sisyphus weep, so I can only imagine how the drivers must feel. Why would anyone condemn them to this fate? The rack of lidar sensors, radar, and cameras on top of each vehicle gives it away. Their owner, a startup down the street called Zoox, believes that if the fleet captures enough data — if it encounters enough wildly different urban anomalies that might pop up on the road (a plastic bag in the air, a bike on the back of a car, a child chasing a ball in the rain) and learns to recognize and plan around them — then boom! The age of self-driving vehicles will have arrived, ushering in the greatest tectonic shift in transportation since the Wright brothers took off at Kitty Hawk.
Zoox is a five-year-old company worth over $3 billion, having raised $800 million from some of the best venture capitalists out there. One early backer, the firm Lux Capital, recently raised a $1 billion fund based on the company’s success as a crown jewel in its portfolio. It’s one of a number of companies — including Aurora, GM’s Cruise, Alphabet’s Waymo, and even Tesla — betting on the super-intelligent Mr. Magoo theory of self-driving cars. Like Mr. Magoo, their vehicles experience technical hitches due to near-sightedness and a stubborn refusal to admit the problem. Their hope is that if they jack up Mr. Magoo’s brain to superintelligent power and have him circle the block a trillion times, his poor eyesight won’t stop him, because the recognition algorithms running on his visual cortex will be able to identify the intentions of another driver from the faint blur of a mere handful of pixels, even at a 150 yard distance while traveling 60 miles an hour.
Good luck with that, especially at night in the rain.
Of all the Magoos, Tesla at least has the advantage of drawing data from 500,000 cars let loose on the roads. As of this spring, Tesla was training and updating its self-driving software from the experiences of 500,000 cars with a radar, eight cameras, and twelve ultrasonics. “All the cars being produced have all the hardware necessary — computer and otherwise — for full self-driving,” Tesla CEO Elon Musk said at his company’s Automation Day presentation on April 22 of this year. “I’ll say that again: All Tesla cars being produced right now have everything necessary for full self-driving. All you need to do is improve the software.”
Full self-driving hardware! Just take Mr. Magoo around the block a few trillion times and he’ll be fine. They’ll update his brain. Of course, as the many videos of Teslas self-parking on YouTube will attest, Musk’s cars park like, well, Mr. Magoo: poorly and slowly. Perhaps failure to live up to Musk’s proclamation explains why 32 Tesla executives have been fired or asked to step down in the 18 months or so.
Musk has to exaggerate because Tesla is a car company whose stock trades like a tech company. Tesla might sell 400,000 cars this year. By contrast, Ford might sell 6 million, GM 8.5 million. Granted, the Tesla Model 3 looks and drives like a dream. But when you count salaries and overhead according to Tesla’s own quarterly statements, it costs more to make a Tesla than people are willing to pay for it. And that calculus includes the federal subsidies that will dry up on December 31 of this year. Ford is worth $35 billion and makes money on its cars. Tesla is worth $40 billion and doesn’t. How is this math possible?
Tesla’s stock trades at such a large multiple of its revenue because Musk has convinced shareholders that it’s not a car company, but an artificial-intelligence company that happens to use a fleet of 500,000 cars to collect and label data. It’s a clever sleight-of-hand, but it’s not fooling those who matter. As a fund manager on Wall Street once told me, “You’re not a hedge-fund manager until you’ve shorted Tesla at least once.”
Well, win some, lose some, wreck some. The venture-capital industry and the autonomous-vehicle market wouldn’t be nearly as exciting if some people didn’t get sacked, sued, go under, blow up, or beg for mercy in Chapter 11 proceedings every so often. For those in the grandstands watching the tech world, it’s lights out and away we go. The industry is one or two hairpin turns away from a colossal crash of bankruptcies, acquisitions, foldings, and consolidation. If the prize is rich and the future a golden promise, there is still a long way to go, and in the short term things are going to get nasty.
In fact, we estimate that ninety percent of the startups in the autonomous-vehicle space today will not exist in five years. A startup called Drive.ai was putt-putting on fumes when Apple scooped it up for pennies in an acquisition this summer. Driving-assist startup Scotty Labs was just acquired by DoorDash. Intel bought the automotive-camera maker MobileEye for $15 billion. Ford and Volkswagen recently announced they were teaming up instead of going at it alone. For the industry itself, this is the beginning. But for many of the players, the end is near.
The big crunch is coming because, over the next year, all the major auto and trucking companies will decide on who will be the suppliers for their main production lines in 2022. This won’t be for full self-driving, but for something a little more modest if still vitally important: a car so safe it is incapable of crashing. Once the major auto companies settle on their suppliers, once the music stops, it’s going to be kaput for anyone else who doesn’t have a chair.
To lay my cards on the table, my own venture-capital company, the 1517 fund, is betting that the Magoos’ labor will continue to yield weak results. Our biggest investment is in Luminar Technologies, a company making a perception system using a wholly unique lidar. Instead of only giving self-driving cars a brain, Luminar wants to give them better-than-human eyes, too. With key improvements on four core components — laser, receiver, microchip, and scanner — Luminar’s lidar is the only one that currently satisfies the major auto companies’ technical requirements for full autonomy in a vehicle. In other words, the data coming into the Luminar system is so rich that recognition algorithms can work faster and more accurately with less time to learn, even at night or in the rain.
Given these advantages over the Magoos, our fund fully expects to have a chair when the music stops. Our bet is that all the genius VCs backing contrarian founders and all the CEOs of companies commanding fleets that drive in circles for years will end up disappointed, while the rest of us building the future without lies and exaggerations celebrate. This is provided, of course, that government bureaucrats aren’t allowed to tip the scales and ruin the fun in the meantime, as I certainly hope they aren’t.
After all, if Elon Musk, Google, Apple, Intel, and the best and brightest VCs in the business could make such a colossal miscalculation, a federal pencil-pusher doesn’t stand a chance.