Uber puts $10bn behind 120,000 robotaxis it will not build - and buys its way into the companies that do

Alongside its second-quarter results in the first week of August 2026, Uber said it would commit more than 10 billion dollars to putting 120,000 autonomous vehicles on its platform. The money splits roughly two ways: about 7.5 billion dollars to buy fleets, released against deployment milestones rather than up front, and about 2.5 billion dollars in equity stakes in the companies that make the cars and the driving software - Lucid, Rivian, Nuro and Wayve among them. The service target is at least 15 cities by the end of 2026 and 28 cities by 2028. What is notable is the shape of the bet rather than its size. Uber has spent years insisting it does not need to build a self-driving car, having sold its own programme in 2020, and this announcement doubles down on that: more than thirty partnerships, no in-house vehicle, and capital used to secure supply rather than technology. The company is buying the two things a robotaxi operator cannot generate on its own - hardware it does not manufacture and priority access to it - while keeping the part it already owns, which is the demand. The underlying quarter was strong enough to pay for it: 208 million monthly active platform consumers, up 16 percent, trips up 18 percent, gross bookings up 22 percent to 58.02 billion dollars and revenue up 11 percent to 14.19 billion. The market response was cooler - the stock has fallen through 2026 - which reads less as doubt about the results than about how much of the value in a driverless network eventually accrues to the app rather than to whoever owns the cars. The number to keep in view is the one Uber gave earlier this month: robotaxis are live in seven cities and still account for well under half a percent of its trips. Going from that to 120,000 vehicles is not an incremental step, and the commitment is deliberately staged so the money follows the deployments rather than leading them.