Uber has made no bones with what it wants: a fleet of self-driving cars on the road, perhaps one day replacing human drivers. It’s spent tremendous amounts of money to do that, hiring hundreds of engineers, establishing the Uber Advanced Technologies Group, and making high-priced acquisitions, like self-driving trucking company Otto for $680 million (an acquisition that has ensnared Uber in a long and costly legal battle.) So what is its biggest competitor Lyft up to?
At a press briefing Thursday, Lyft opened up about its desire to be an “open platform” for self-driving-car providers like Waymo, nuTonomy, General Motors, Jaguar Land Rover, and more. The idea is, Lyft would consolidate passengers’ requests, with Lyft then determining if the ride request should be handled by an autonomous vehicle (or — in the case of inclement weather, road construction, or a somewhat unknown route — a human driver). It would be an agnostic platform — your ride might be provided by Waymo or by General Motors — but all handled through the Lyft app.
More surprisingly, Lyft will also be building out its own self-driving efforts. It will be opening a facility in Palo Alto that it’s dubbing Level 5, named after the most advanced form of self-driving-car tech in which no human interaction is required. Lyft says it plans to have hundreds of employees working at the Palo Alto facility by the end of 2018.
Lyft execs envision a world where fewer and fewer people own cars, and transportation is increasingly viewed as a “subscription.” Lyft’s chief strategy officer, Raj Kapoor, trotted out some now-familiar numbers: The average American spends $9,000 a year on their vehicle, but only uses it four percent of the time, and the other seats in a car are empty for 80 percent of rides.
So why would companies like Waymo and nuTonomy, or General Motors and Jaguar Land Rover, normally competitors, want to enter into a partnership with Lyft where information may get freely shared around? Lyft’s director of product, Taggart Matthiesen, says everyone in the field is recognizing the challenge in front of them. “The world has changed from a couple years,” he said. “People have realized how difficult, how expensive it is.”
To sweeten the deal, Lyft will be building software and hardware for cars in its network at its Level 5 facility. It’s unclear what exactly this tech will do, but one example offered was simply providing data. “The way that these algorithms improve in new situations,” said Lyft’s Kapoor, “the big factor is, how many new situations is it seeing?”
Lyft now completes 1 million rides per day, covering tens of millions of road miles. If Lyft was able to install a sensor package on the majority of its cars and then share that data with its partners focused on self-driving cars, it could be enough to get partners who might otherwise be leery to join in.
Lyft — which fights with Uber over recruiting and retaining drivers for its network — tried to reassure its current drivers who may be leery about being replaced by an AI. At the press briefing, a spokesperson said the number of Lyft drivers “would never go lower than it is today,” while a blog post by vice-president of engineering Luc Vincent affirmed that Lyft “will always operate a hybrid network, with rides from both human-driven and self-driving cars.”
Lyft also affirmed that it was not planning to get into the business of manufacturing self-driving cars. That means details would need to be hammered out about the revenue split between it and its partner companies, while also keeping prices at a point where consumers see ride-hailing as a part of their daily life, rather than an occasional extravagance.
There’s a lot that remains unclear about Lyft’s efforts in self-driving cars. Still, there’s little question of why it’s doing it. As Lyft VP Luc Vincent put it during the press briefing: “It’s too strategic an area for us not to be a player.”