:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/63711527/152766139.0.jpg)
When Google quietly began working on cars that drive themselves, über was just a German word.
Now, nine years later, Uber the ride-hailing company looks set to get regular people inside autonomous vehicles first — a move that’s critical for Uber, and dispiriting for the audacious project hatched inside of Google.
Uber said on Thursday that, along with Volvo, it will test a fleet of on-demand autonomous cars in Pittsburgh later this month. An Uber driver will still be in the car. But, more importantly, so will a customer.
That’s something that Google has yet to do.
When the search giant first unfurled its self-driving plans, it was the only company tackling the advanced tech. Not anymore: All the major car companies have joined the fray — and shown a willingness to cut big checks for the tech behind it.
Google is well ahead on the technical challenges of driverless vehicles, according to most in the industry. But those same people, in recent months, have begun asking why, nearly a decade after hiring top roboticists to build its project, Google has not delivered something to market.
There are a number of reasons why Google has been a laggard to new rivals in the field.
One is its devotion to going fully driverless, a far more difficult feat. Google’s cars have driven over 1.8 million miles with test drivers in the front seat. Those test drivers could carry passengers (barring any legal obstacles), but don’t.
That’s partially because, unlike the car companies and Uber, the Google self-driving cars don’t need an immediate revenue stream. They have search ads to bankroll them.
Uber has a more pressing need to develop self-driving cars. CEO Travis Kalanick told Bloomberg the tech was “basically existential for us.”
Yet other forces may be holding back Google’s cars — like internal dynamics at the project, now within the X subsidiary under Google parent Alphabet. Alphabet has kept its revenue strategy for the unit under wraps. Some sources say that is because they have not settled on one yet.
“They went back and forth all the time,” said one person who recently left X.
A rep for the X project declined to comment.
Finding partners is also a factor.
The X project has announced a small deal with Chrysler, but sources say that most U.S. auto companies have balked at cutting deals, wary of the data and control the Google-y unit would demand. That may be why several sources have said the X team has spoken to Chinese manufacturers lately.
“They’re more willing to bend steel,” said one investor in the industry.
Google still has an edge over Uber and car companies in machine intelligence. Google has better maps. (Although would-be rivals like Uber and Apple, which is working on something car-related, and the car makers are scrambling to improve their own maps.)
And Google has Android — a capacity to distribute software for self-driving cars on half the smartphones in the U.S. and 80 percent of them worldwide.
Still, frustration with the inertia of Google’s self-driving car looks to be hitting its own ranks.
Chris Urmson, the CTO and former director of the car unit, recently departed, along with two early engineers.
Earlier this year, several members of the project, including co-founder Anthony Levandowski, decamped to form Otto, an autonomous trucking startup. With its announcement this morning, Uber also said it had acquired Otto.
This article originally appeared on Recode.net.
