Conventional wisdom on self-driving used to go like this: A smart tech company, like Google's Waymo, writes the self-driving software. A smart chip company, like Nvidia, provides the computing power. A smart automotive supplier, like Delphi, brings the car parts. A smart carmaker, like GM, furnishes the cars—and everything runs on the platform created by a smart ride-hailing company, like Uber or Lyft. This line of thinking explains why players in this industry are partnering up like their survival depends on it. In the stage production that is getting you from A to B in a robocar, everyone plays a role.
But some of these companies want bigger parts than others. Today, Lyft announced it’s getting into the self-driving business, launching its own unit to build autonomous vehicle software and hardware. “It’s too strategic an area for us to not be a player,” says Luc Vincent, the ex-Google StreetView-er who will head up the initiative's technical side.
Until today, Lyft's strategy seemed to hinge on hopping between carmakers like General Motors and tech companies like Waymo, striking deals that would put autonomous vehicles on the Lyft platform. (The company did say it'll maintain these partnerships, including one with self-driving car company Nutonomy that's slated to bring self-driving taxis to Boston.) But Lyft's strategic play to tackle autonomy on its own suggests that business models for this booming industry are far from settled.
Lyft is entering the self-driving race an underdog’s underdog. Waymo has been testing self-driving vehicles in the Bay Area since 2010, collecting more than 3 million miles' worth of data on public roads and the vehicles that traverse them. Arch-rival Uber poached top-notch robotics talent from Carnegie Mellon University to create its self-driving unit back in 2015.
And dedicating hard-won money to self-driving tech is riskier for a company like Lyft than an auto-making incumbent. Just look at the money those other companies have to burn: $19.8 billion for Ford, $16.6 billion for General Motors. Lyft is valued at $7.5 billion. "Not only is does General Motors have cash on hand, but its cost of capital is a lot lower than Lyft’s," says Erik Gordon, who studies tech entrepreneurship at the University of Michigan’s Ross School of Business. The biggies have less to lose.
Lyft's move might be defensive. "I think what Lyft is worried about is envelopment," says Evan Rawley, who teaches entrepreneurial strategy at the University of Minnesota, Twin Cities. "Say Waymo becomes the dominant autonomous vehicle company. Then it’s very easy for Google to attach a ride-hailing service to that and make it hard for Lyft to compete in their core business."
The move is similar to what Google did with Android, he says. The search giant didn't want Apple to own everything related to smartphones: the hardware, the software, all the nutty stuff that goes on in the App Store. So it built Android, an operating system, and kept its hardware cheap, hoping to ensure their role in the smartphone universe even if Apple kept on killing it with the iPhone. Now lots of hardware companies use Android as their operating systems, and Google phones are still around. Lyft, like Google, is hedging its bets.
The worst case scenario is that Lyft gets stuck selling to one, monopolistic self-driving giant. The company wants to keep the autonomous vehicle space a little messy, and wants lots of friends with whom to horse trade. "We want to take a proactive approach in pushing the industry into more of an open environment," says Raj Kapoor, Lyft's chief strategy officer. Now it can offer partners years of knowledge about how passengers behave, and some gadgets to help the cars get around. Bolt it on and go.
No one's quite sure how this self-driving thing will shake out. It's not even clear whether companies like Lyft and Uber will have an edge, once human drivers are out of the ride-hail equation. "Maybe the value creation isn't going to be in the ride-sharing platform," Gordon says. Maybe building software will be the only way to stay involved in a fast-changing, self-driving market.
Now, Lyft does have some good stuff going for it. The company wan't say how much it's dedicating to self-driving, but it will hire hundreds more engineers to work on it, from a new office in Palo Alto. (It's called the "Level 5" building, Level 5 being engineer-speak for full autonomy, natch.) It also claims to have 1.2-billion miles' worth of data on traffic flow, travel times, and where passengers want to be, and when. Oh, and the company says it will make use of its self-driving network to train its software, mounting sensors on vehicles drivers lease through its Express Drive program. It might also offer drivers incentives to stick their own cars with data-collecting sensors, says Vincent, the Lyft engineer. Driving themselves out of a job? Sure. But Lyft also maintains that humans and robots will coexist on our roads for at least thirty years, the fleshy beings substituting where the self-driving cars cannot.
While that engineering work speeds on, expect more self-driving partnerships from Lyft. It's a dance or be danced world out there, and no one's quite sure when the music stops.
