Uber is Giving Up Self-Driving Cars (and Flying Ones, Too)

By John Banks

For Travis Kalanick, the prior CEO of Uber, self-driving cars were an inevitable destiny. For Dara Khosrowshahi, Uber’s current CEO, they are an expensive distraction to get rid of.

Uber has a five-year-old self-driving car unit known as Advanced Technologies Group (ATG). The unit has roughly 1,200 employees, many of them in Pittsburgh, Pennsylvania (a self-driving mecca due to the long-time robotics focus of Carnegie Mellon University).

In 2019, Uber’s self-driving unit raised capital as a standalone entity with a valuation of $7.25 billion. But on Dec. 7, Advanced Technologies Group was revalued lower, at just $4 billion, as part of a deal that spun the whole thing off to Aurora, a self-driving competitor that is valued at $10 billion.

Getting rid of its self-driving unit means that Uber can stop burning capital and resources in that area, and focus more on rideshare and food delivery instead (its two main lines of business).

At the same time, Uber will invest $400 million in Aurora and have a 26% stake, which Bloomberg calculates is closer to 40% when the personal stakes of Uber employees and investors are added in.

The close ties between Uber and Aurora mean that, when self-driving cars start to roll out in a serious way, Aurora will have full access to Uber’s rideshare network in a partnership that benefits both.

And yet, Travis Kalanick’s original worry still seems valid.

Kalanick believed that, eventually, the rideshare business would convert to a 100% self-driving model because of the economics, in a world where human drivers were obsolete and car ownership was rare.

“The reason Uber could be expensive is because you’re not just paying for the car — you’re paying for the other dude in the car,” Kalanick said at a technology conference in 2014.

“When there’s no other dude in the car, the cost of taking an Uber anywhere becomes cheaper than owning a vehicle,” he added. “So the magic there is, you basically bring the cost below the cost of ownership for everybody, and then car ownership goes away.”

Six years later, in 2020, the self-driving-car revolution is taking far longer than Silicon Valley expected.

This explains why Uber’s current CEO is exiting the self-driving business, while still maintaining an option to participate via the Aurora spin-off stake if and when self-driving takes off as a business.

It remains a possibility that Uber never becomes profitable, ever. The company is expected to lose at least a billion dollars in 2020, if not more, continuing a long unbroken string of eye-bulging losses. The company’s valuation of $93 billion, as of this writing, is based entirely on the hopeful assumption that the company will figure it out one day.

Without the benefits of being a large-scale, self-driving player, obtaining profitability will become that much harder for Uber (and Lyft, too).

The problem comes down to the terrible economics of human-driver-based rideshare and delivery businesses, coupled with brutal competition. The only way to make driver-based businesses work at scale, so far, appears to be forcing a low-wage existence onto the drivers, while getting them to fork out gas, vehicle, and insurance costs from their own pocket.

The model could theoretically work if, say, a competitor like Uber could grow large and dominant enough to have a near-monopoly in all geographical regions where it competes, and then raise prices enough to turn a profit and help drivers make a real living. 

But that isn’t going to happen anytime soon, because cutthroat competition among rideshare and delivery services is still going strong.

As an example of this, the delivery service DoorDash is expected to go public this week at a valuation near $40 billion. Grubhub, another food delivery competitor, has a market cap of just over $6 billion.

And there is also this little company you may have heard of, called Amazon, that already has such an extensive network of trucks and drivers it could conceivably dominate delivery service with the flick of a switch.

So, Uber is getting out of the self-driving car business — which requires large amounts of capital and investment for an uncertain payoff, years down the road — in order to focus on its food and rideshare delivery businesses, which have terrible economics and still face self-driving as an existential threat. 

Meanwhile the most fearsome competitor in the self-driving space is probably Waymo, the self-driving car company that began as a unit of Google and was valued in March at $30 billion.

In October of this year, Waymo announced a partnership with Daimler Trucks — the largest commercial vehicle maker in the world by volume — to enable self-driving big rigs.

The Waymo-Daimler partnership goal was to create “a full suite of driving capabilities,” said Waymo’s CEO, including “not just highway driving but full hub to hub capability, including the ability to navigate very complicated city environments.”

Waymo is also building a full-blown artificial city in East Liberty, Ohio, to test self-driving cars under realistic urban conditions. The Ohio location means Waymo’s cars will also gain experience with snow and ice conditions, to pad out the knowledge it has gained from more than six million miles’ worth of driverless miles logged in Phoenix, Arizona.

The global auto industry is projected to be worth nearly $9 trillion dollars by 2030. That is a staggeringly large ecosystem in terms of not just software, but physical technology.

Think of all the steel and fiberglass, cameras and sensors, vehicle and roadway modifications, and communication network hook-ups, on and on, that will have to be rolled out globally to enable autonomous self-driving capability on a mass scale.

In the face of all that physical infrastructure, and likely costs running into the hundreds of billions, Waymo’s long-term bet is that partnering with existing automakers (like Daimler) is the way to go, so that Waymo can focus on the highest value-add part of the system — the technology itself — and leave others to build the vehicles (at a significantly higher investment cost, and lower profit margin). 

The Waymo approach makes the most sense to us. If we had to pick a winner in the ultimate self-driving race, it would be Waymo.

  • Waymo is building its approach around partnerships, which means their self-driving tech could scale faster across existing auto fleets in the future.
  • By building connections within the global auto industry across a range of original equipment manufacturers, Waymo will have the ability to harness competition in favor of its business model, rather than trying to take it head on. 
  • It doesn’t hurt that Waymo’s deep-pocketed parent is Alphabet (the owner of Google), that their valuation of $30 billion (likely higher by now) already puts them ahead, and that they are focused on the long game (even if profitability takes another five years).

As for Uber, spinning out their self-driving unit seems like a smart tactic, at least in the near-term. It is a way to placate investors who are anxious to see Uber reach profitability via rideshare and food delivery alone — a feat that might be akin to a miracle, but which plenty of investors still think is possible (as evidenced by Uber’s wacky share price).

And so, to the extent self-driving profits are too far-off a prize to justifiably keep shoveling money into that furnace, getting rid of the self-driving unit is a good thing. Unfortunately for Uber, this form of smart management is reminiscent of what is going on at IBM, where savvy financial engineering movements are akin to rearranging the deck chairs on the Titanic (in terms of actual long-term strategy and the state of the core business).

Last but not least, in a showy display of focus — the CEO may as well be shouting, “Look how disciplined we are being!” — Uber is giving up on its flying car ambitions, too.

“Uber is handing its flying car project, Uber Elevate, to the air taxi start-up Joby Aviation,” the New York Times reported a day after the self-driving spin-off news.

“Uber will also invest $75 million in Joby’s effort to build a flying taxi,” the NYT added, “while agreeing to become partners with the start-up when the flying car reaches the market.”

Which will come first, we wonder: Uber reaching actual profitability (after a decade-plus of routinely burning billions in business lines that still look terrible), or the Joby flying car debut?

Call it a toss-up, with “neither” the favored bet.