Like a parent snatching away the car keys, the California DMV has put the kibosh on Uber's autonomous vehicle trial in San Francisco.
On Wednesday, an Uber spokesperson said, California stripped the registration from the autonomous Volvos that the company had put into service in San Francisco. The trial has ended, though the company says it will find a way to "develop workable statewide rules" in California going forward. It will seek to deploy its autonomous vehicles elsewhere in the interim.
It's a humbling moment for the San Francisco–based ride-hailing giant, which had maintained that its autonomous vehicles did not require a special permit under state law. They were not actually autonomous because they could not drive without human monitoring, the company's AV czar Anthony Levandowski argued in a conference call with reporters last Friday. He did not see what distinguished Uber's cars from Teslas with autopilot enabled.
But it's also a crucial test case. Uber has long flouted local laws until they could be bent to suit its operations or pre-empted in the statehouse. As the San Francisco case demonstrates, that posture gets harder to maintain as the company shifts from software to hardware. The days of Uber playing the outlaw startup are vanishing as the company acquires millions in fixed assets.
The weeklong trial had been contentious from the start. The cars, which are already in customer operation in Pittsburgh, were filmed and photographed running red lights, though the company blamed that on human error. Several days later, the San Francisco Bicycle Coalition alleged the cars were making unsafe turns across bike lanes. The company said engineers were working on the problem.
That Uber proceeded with the rollout despite assurances from California regulators that it would be shut down demonstrates the company's hubris, you might say. But it also shows the company's determination to evolve from a software company to a fleet manager. Developing an autonomous vehicle is "basically existential for us," CEO Travis Kalanick told Bloomberg this summer.
The latest financials show why. Uber lost $800 million in the third quarter of 2016, investors told the Information, a tech news site, due to the company’s heavily subsidized operations. To make its way to profitability, Uber will either have to raise ride costs significantly or develop autonomous vehicles, which are projected to cut per-mile ride costs by as much as 75 percent.
Moving from a revolutionary piece of software into the development, ownership, and maintenance of a massive vehicle fleet will be quite the corporate pivot. In Madrid, the company launched a fleet of Teslas on Thursday. The idea is to use green cars to soften resistance to Uber in the Spanish capital, which suffers from air pollution produced by congestion. But it is also another example of Uber's expanding profile of hard assets, and an experiment for the challenges that come along. Where will the cars be parked? Who will handle maintenance? What kind of wear and tear will they sustain?
What San Francisco illustrates is how actually having cars on the road works to constrain Uber's corporate approach—and emboldens regulators. Uber broke the power of the municipal taxi cartels by breaking the law, and in most cases, cities found ways to legalize it—and suffered if they fought back. When a city passed what the company considered debilitating regulations, Uber walked. This cavalier move burned the company's drivers, but Uber, unlike a company like Carrier in the midst of government negotiations, could reopen its business with the click of a button. Literally.
Unlike factories, cars can be moved to a different city, of course. But somehow the prospect of a thousand Ubers caravaning from Madrid to Valencia seems like an unlikely sequel to the company's earlier flights from regulation. It's hard to be nimble when you're dragging around several millions of dollars of cars, depreciating every minute they're not in use.