Honestly, if you were scrolling through tech news back in 2016, you probably thought we’d all be napping in the backseat of a Volvo by now while a robot navigated rush hour. Uber was the loudest voice in that room. Travis Kalanick, Uber’s former CEO, famously called the quest for autonomous tech "existential." He wasn't exaggerating. To him, the human driver was the biggest expense on the balance sheet. If Uber didn't figure out self-driving cars, he feared Google or Tesla would simply delete them from the market.
Fast forward to today.
The story of self driving cars Uber isn't just about a company trying to save money on labor. It’s a messy, high-stakes drama involving a trade secret lawsuit with Waymo, a tragic pedestrian fatality in Arizona, and a massive corporate pivot that saw Uber literally pay another company to take their autonomous division off their hands.
The Rise and Fall of the Advanced Technologies Group
Uber didn't start small. They went straight for the throat by poaching dozens of researchers from Carnegie Mellon University’s National Robotics Engineering Center in 2015. It was a talent raid that practically gutted one of the most prestigious robotics programs in the world. Suddenly, Pittsburgh became the hub for Uber's "Advanced Technologies Group" (ATG).
They were burning cash. Fast.
At one point, Uber was spending roughly $20 million a month on R&D for autonomous systems. The goal was a "Level 5" vehicle—one that could go anywhere, anytime, without a steering wheel. But the reality on the ground was much crunchier. The cars struggled with simple things like unprotected left turns or "edge cases" like a cyclist carrying a large mirror.
Then came the legal firestorm. Anthony Levandowski, a former star engineer at Google’s self-driving project (now Waymo), joined Uber after his own startup, Otto, was acquired for hundreds of millions. Waymo sued. They claimed Levandowski had downloaded 14,000 highly confidential files before leaving. The lawsuit was a PR nightmare. It eventually settled, with Uber giving Waymo about $245 million in equity, but the momentum was fractured.
The Tempe Tragedy and the Shift in Public Trust
Everything changed in March 2018. In Tempe, Arizona, a self-driving Uber test vehicle struck and killed Elaine Herzberg as she crossed a road at night with her bicycle. It was the first recorded pedestrian death involving a fully autonomous vehicle.
The investigation by the National Transportation Safety Board (NTSB) was damning. It revealed that the software had detected Herzberg about six seconds before impact but struggled to classify her—first as an unknown object, then a vehicle, then a bicycle. The system's emergency braking had been disabled to prevent "erratic driving" behavior, relying instead on a human safety driver who, dashcam footage showed, was looking down at a phone.
Uber grounded its fleet immediately.
This wasn't just a technical failure; it was a cultural one. The "move fast and break things" mantra of Silicon Valley doesn't work when "things" are human lives. The company eventually returned to the streets with much stricter protocols, but the magic was gone. The public was skeptical. Investors were tired of the "money pit" reputation of ATG.
Selling the Dream to Aurora Innovation
By 2020, the world looked different. The pandemic was hitting ride-sharing hard, and Uber needed to show a path to profitability to satisfy Wall Street. Dara Khosrowshahi, who took over as CEO after Kalanick, started stripping the company back to its core: moving people and moving food.
In a move that surprised many, Uber sold ATG to Aurora Innovation, a startup led by Chris Urmson, another former Google engineer. But it wasn't a standard sale. Uber actually invested $400 million into Aurora and handed over the entire division. Basically, they paid Aurora to take the self-driving burden off their plate while retaining a 26% stake in the company.
It was a total white flag on building the hardware themselves.
Where Are the Uber Self Driving Cars Now?
If you open the Uber app in certain cities today, you can actually catch a ride in a car with nobody behind the wheel. But it’s not an "Uber" car.
✨ Don't miss: Apple TV APK for Android: What Most People Get Wrong
Uber has pivoted from being a developer to being a platform. They realized that building the "brain" of a robot is incredibly hard, but managing the "marketplace" is what they do best. Today, Uber partners with the very companies it once tried to beat:
- Waymo: In Phoenix, you can specifically request a Waymo through the Uber app. It’s the ultimate "if you can't beat 'em, join 'em" scenario.
- Motional: Uber signed a 10-year deal with Motional (a joint venture between Hyundai and Aptiv) to deploy autonomous taxis and delivery robots.
- Waabi and Aurora: These partnerships focus more on the freight side, with Uber Freight integrating autonomous trucking tech into its logistics network.
The strategy is now "asset-light." Uber provides the customers and the routing software; companies like Waymo provide the expensive, sensor-laden hardware.
The Reality Check: Why This is Harder Than We Thought
We used to hear that self-driving cars were "two years away." We've been hearing that for a decade. The reason Uber's original project struggled—and why even the leaders are moving slowly—comes down to the "long tail" of probability.
Driving is a social act. It’s about eye contact with a pedestrian to signal they can cross. It’s about understanding that a ball rolling into the street usually means a child is following it. Computers are great at following rules, but they are historically terrible at nuance.
Rain, snow, and even heavy dust can blind LiDAR sensors. Construction zones with hand-signaling workers confuse the algorithms. Uber's early tech was great in the sunny, predictable grid of Phoenix, but it struggled in the chaotic, narrow streets of Pittsburgh or the fog of San Francisco.
What This Means for the Future of Work
The anxiety about robot cars replacing drivers hasn't vanished; it has just shifted. For the millions of people who drive for Uber globally, the "threat" is no longer a fleet of Uber-branded Volvos. It’s a slow, city-by-city integration of third-party autonomous vehicles into the app.
It won't happen overnight. We are looking at a hybrid future where autonomous cars handle the easy, high-traffic routes in fair weather, while human drivers handle the complex trips, the grocery hauls, and the rural areas.
Actionable Insights for the Tech-Curious
If you're following the trajectory of self driving cars Uber and the broader industry, there are a few things you can do to stay ahead of the curve:
- Watch the Partnerships, Not the Hardware: Don't look for "Uber cars." Look for who Uber is signing deals with. The winner of the autonomous race will be the company that can integrate into existing apps seamlessly.
- Check the Geo-Fencing: Autonomous rides are currently limited to "geo-fenced" areas. If you live in Phoenix, San Francisco, or Las Vegas, you can test this tech today. If you're elsewhere, it's still years away.
- Monitor Uber Freight: The first place you'll likely see a massive economic impact isn't in ride-hailing; it's in long-haul trucking. Uber is heavily betting on autonomous trucks to solve the driver shortage in logistics.
- Safety Data is Key: Before getting in a driverless car, look at the disengagement reports. Companies like Waymo and Cruise (owned by GM) publish data on how often a human has to intervene. It's the only real metric that matters.
The dream of the Uber self-driving car isn't dead—it just got a massive reality check. The company that once wanted to own the robots has decided it's much safer, and cheaper, to just be the one who introduces you to them.