UberUBER

pressed the eject button on self-driving vehicles back in 2020 with the sale of its autonomous business unit, but now it’s frantically trying to rebuild as consumers embrace a world without drivers. To catch up, the ride-sharing and delivery app has set ambitious targets and announced major investments and partnerships to transform itself from a marketplace for human drivers into a robot fleet commander.

Shifting gears: McKinsey projects robotaxis will become commercially available at scale in the US by 2030. While Waymo already operates over 2K electric robotaxis delivering 250K trips weekly across major US cities, Uber is falling behind and now it’s ramping up partnerships, including recently announced plans to operate a fleet of 100K self-driving vehicles powered by Nvidia’sNVDA

technology. As part of this deal, StellantisSTLA

committed to delivering at least 5K Nvidia-powered robotaxis starting in 2028.

• Uber has also struck deals with more than a dozen robotaxi developers, including a commitment to purchase 20K vehicles from Lucid and Nuro over six years.
• The company is also investing up to $375M into Avride, aiming to expand the Dutch subsidiary’s driverless fleet to ~500 cars while supporting product development across new markets.

Uber Stays in Neutral

The company is positioning itself as the “Switzerland of self-driving vehicles” by integrating self-driving vehicles from various manufacturers. While this could help Uber unlock massive cost savings — with estimates suggesting that self-driving vehicles could reduce ride-hailing costs by half — some competitors have plans to cut off Uber entirely. Waymo and Tesla have already built consumer apps for users to hail rides, and both companies have their advantages over Uber — Waymo being owned by deep-pocketed AlphabetGOOG

and TeslaTSLA

having strong hardware expertise.

• Instead, Uber is hoping its massive customer base and operational expertise will keep it relevant, pointing to how Starbucks sells through both its own app and Uber Eats despite the competition.
• Luckily for Uber, Elon Musk is struggling with its launch, having recently scaled back Tesla’s robotaxi rollout, shifting from plans to be in “half the US population by the end of 2025” down to launching in just 8-10 US metros over the next two months.

The cost issue: Uber CEO Dara Khosrowshahi identified hardware costs as the biggest barrier to widespread adoption. Current autonomous vehicles packed with sensors, cameras, lidar, and computing power often exceed $100K per unit — making profitability elusive. The CEO predicts it’ll take ~5-6 years to bring costs below that threshold in the US market as Chinese competitors like Baidu’s Apollo are already producing cars for under $50K. And if Uber fails, it’ll feel the same disruption it once inflicted on the taxi business.