Aurora, one of the nation’s leading self-driving startups, will become the new owner of Uber’s self-driving division, Aurora announced on Monday. In addition to turning over Uber’s self-driving division, known as the Uber Advanced Technology Group (ATG), Uber will also pump $400 million into Aurora.

In exchange, Uber will get a minority stake in Aurora and Uber CEO Dara Khosrowshahi will get a seat on Aurora’s board.

The deal allows Uber to unload a self-driving division that has struggled to regain its footing ever since an Uber ATG vehicle struck and killed a pedestrian in March 2018. Uber shut down its on-road testing for several months after that incident, and the program has faced lingering public skepticism ever since. It’s not clear if the deal will lead to layoffs at Uber ATG.

The deal lets Aurora hedge its bets

Aurora CEO Chris Urmson co-founded Aurora in 2016 after leaving his role as the head of Google’s self-driving project. He assembled a “dream team” along with former leaders at Tesla and Uber’s self-driving projects. The trio built a company that’s widely respected for its engineering rigor—though so far the company hasn’t released much information about the performance of its self-driving technology.

Aurora’s original business model was to license its self-driving technology to automakers. But in recent years, more and more OEMs have taken stakes in other self-driving companies, leaving Aurora with few potential partners in the auto industry. Aurora suffered a setback last year when Volkswagen ended a partnership intended to eventually put Aurora technology in Volkswagen vehicles.

So later in 2019, Aurora pivoted to long-haul trucking as the first application for its self-driving technology. This business model could enable Aurora to succeed without striking big deals with automakers. But it’s risky, since no one knows how soon self-driving technology will be ready for high-speed highway use.

The Uber deal gives Aurora a boost on a couple of fronts. The deal comes with $400 million in cash, which extends Aurora’s “runway” to reach profitability. Uber’s stake in Aurora also increases the likelihood that Uber will be willing to put up more cash in the future if Aurora needs it.

The deal also allows Aurora to hedge its bet on long-haul trucking. Uber ATG’s focus has been on driverless taxis. In a blog post, Aurora CEO Chris Urmson wrote that the deal “gives Aurora an unprecedented opportunity to lead the industry in both trucking and passenger mobility.” Not only does the Uber ATG team bring valuable expertise in the taxi market, but Uber will be a valuable partner when Aurora is eventually ready to deploy driverless taxis across the country and around the world.

A few years ago, Uber faced accusations—and a lawsuit—that it tried to steal lidar technology from Waymo. Aurora acquired lidar startup Blackmore last year, so Aurora can now supply lidar technology to the former Uber project going forward.

Uber doesn’t need to own self-driving technology

An important subtext of the deal is the realization that ownership of self-driving technology probably isn’t the strategic necessity for Uber it seemed to be a few years ago.

When Uber launched its self-driving program in 2015, CEO Travis Kalanick viewed it as a matter of survival for his company.

“If we are not tied for first, then the person who is in first, or the entity that’s in first, then rolls out a ride-sharing network that is far cheaper or far higher quality than Uber’s, then Uber is no longer a thing,” Kalanick told Business Insider in a 2016 interview.

At the time, hype over self-driving technology was nearing its crescendo. Google had unveiled its self-driving technology a couple of years earlier, and many industry observers expected self-driving technology to be widely available by the early 2020s. And so Uber began spending millions of dollars in an effort to catch up to Google.

Uber’s effort was so rushed that it failed to take adequate safety precautions, leading to a fatal crash in 2018 that killed pedestrian Elaine Herzberg. And in retrospect, it’s not clear that Uber’s self-driving project made all that much progress in its first two years. While Uber logged a lot of testing miles before March 2018—second only to Waymo in the US—the cars’ driving behavior was sometimes erratic.

“Every other day in February”

“A car was damaged nearly every other day in February,” one Uber staffer wrote in a scathing internal email in February 2018, shortly before Herzberg’s death. “We shouldn’t be hitting things every 15,000 miles.”

The 2018 crash occurred in part because Uber had disabled some of the vehicles’ emergency braking systems. Apparently the systems had been engaging too often, causing uncomfortable rides. Instead, the company relied on safety drivers to intervene if the car made a mistake.

The resulting crash forced Uber to halt all on-road testing for a few months, then resumed testing on a much smaller scale. The program never regained the sense of momentum it had prior to the crash.

At the time of the crash, I urged Uber to sell the division—perhaps to a major automaker. I argued that giving the project a new corporate parent with a reputation for coloring inside the lines could help to restore public trust. And I noted that Uber was far from turning a profit and didn’t have unlimited cash to throw into R&D projects.

Most importantly, I argued that Uber’s survival didn’t depend on owning self-driving technology. I noted that even after Waymo or others invented a fully self-driving car, it would be slow and expensive for them to build a global ride-hailing network from scratch. So Uber would have a lot to offer companies looking to deploy their self-driving technology on a large scale.

Two and a half years later, I think this analysis holds up well. Self-driving cars have taken longer to come to market than optimists predicted. It will likely take many more years for the technology to become ubiquitous. Being a minority stakeholder in Aurora gives Uber a potential driverless technology partner—without forcing Uber to shoulder the full R&D costs of developing self-driving technology.