Weekly Brief: Another Robo-Taxi Slips Back into the Land of Dreams

Commentary By Andrew Tolve

---

3rd May 2021

Six months after Uber abandoned its self-driving car efforts, rival Lyft is following suit.

The ride-hailing company announced last week that it will sell its self-driving car business, Level 5, to Toyota for $550M. Toyota will pay Lyft $200M upfront and the remaining $350M over the course of five years, by way of a subsidiary named Woven Planet Holdings. The deal is expected to close in the third quarter of 2021.

Lyft’s co-founder and CEO, Logan Paul, framed the sale as “the next phase of an incredible journey to bring our mission to life”. Part of the deal is that if, or when, Toyota commercializes the self-driving technology, it agrees to use Lyft’s platform and fleet data to run its robo-taxi operation.

Despite this upside, the sale of Level 5 marks an ignominious end to the American ride-hailing industry’s foray into autonomous tech development. Back in 2016, when Uber acquired a flashy start-up named Otto for $680M, self-driving cars were viewed as the savior of the ride-hailing industry. The technology had the potential to eliminate ride-hailing companies’ single largest expense, the cost of human drivers, and thrust unicorns like Uber and Lyft into the land of outlandish profitability.

Lyft launched Level 5 in 2017, a year after Uber purchased Otto. It built a 50,000-square foot facility in Silicon Valley, manned it with several hundred engineers and made the audacious claim that come 2021, most Lyft rides would happen in self-driving cars. Instead of hitting that milestone, Level 5 no longer exists and Lyft has given up all internal efforts to pursue self-driving cars.

What went wrong? Covid-19 is one culprit. Lyft lost $1.8Bn in 2020 in part owing to being shut down for months while the pandemic first raged through the United States. Those losses made it harder to justify investing in a cash-intensive proposition like self-driving cars, especially now that Lyft is a public company.

Yet, even without the pandemic, this outcome was probably inevitable. Producing self-driving cars at a level of performance necessary for full deployment has turned out to be far more difficult than even the most brilliant engineers on the planet expected five years ago. Last week, on a quarterly earnings call, Elon Musk said developing self-driving cars has turned out to be “one of the hardest technical problems that exists – that has maybe ever existed”.

That’s coming from a man who wants to bore through the Earth to create hyperloop underground transportation tunnels and who single-handedly reinvigorated the space industry with his SpaceX startup, whose Crew Dragon spaceship completed its six-month flight mission last week when its capsule blazed through the atmosphere and safely splashed into the Gulf of Mexico. When Musk admits that something’s hard, you know it’s hard.

This comes a week after a fiery crash in Texas, where two men died in a Tesla Model S and set off a federal investigation. Authorities believe that no one was behind the wheel at the time of the accident. Musk pushed back on those claims, arguing that data logs suggest Autopilot had not been engaged. Last week on the earnings call, Tesla vice-president of vehicle engineering, Lars Moravy, acknowledged that one component of Autopilot had indeed been engaged at the time of the accident. That component was adaptive cruise control. He also argued that further evidence collected at the scene clearly suggests that there was a driver behind the wheel.

A lot remains to be discovered. Either way, the fact is that developing and deploying self-driving cars is expensive, time-consuming and fraught with risk. From that perspective, Lyft selling off its Level 5 business makes good sense. The company gets an immediate injection of cash, eliminates the expense of future R&D, avoids potential negative press from accidents during development and has a guaranteed network for its technology if Toyota’s service ever comes to fruition.