Uber taps Rivian to build robotaxis in deal worth up to $1.25B

When Rivian Decided It Wanted to Be a Robotaxi Company

There is a version of this story where it makes perfect sense. Uber has a platform, not a factory. Rivian has a factory — one in Normal, Illinois, a second under construction in Georgia — and a CEO who has spent the last four years trying to turn an electric truck company into something broader. Put them together, add 50,000 R2 SUVs, and you have a robotaxi fleet that neither company has to fully own.

That is the deal Rivian and Uber announced on Thursday. Uber is putting in an initial $300 million investment in Rivian and expects to purchase 10,000 fully autonomous R2 robotaxis, with an option to buy up to 40,000 more starting in 2030. The two companies said they plan to launch in San Francisco and Miami in 2028, expanding to 25 cities in the U.S., Canada, and Europe by the end of 2031. The fleet will operate exclusively on Uber's network.

The financial ceiling of the deal is $1.25 billion — a number that sounds significant until you consider that Rivian posted a net loss of $3.626 billion in 2025, per its FY2025 8-K filed February 12, 2026, and has never turned a profit. This is a bet, not a bailout.

What makes the bet legible is the person making it. RJ Scaringe, Rivian's founder and CEO, has made automated driving technology the company's stated top priority — ahead of trucks, ahead of consumer EVs, ahead of everything except not losing money. At SXSW last week, he was blunt about where the money was going. "Our path to get to hands-off, eyes-off in 2027 is something we're spending more money on than anything else," he said.

That sentence contains the entire promise and the entire problem.

Rivian hasn't started producing the R2 SUV yet. Manufacturing is supposed to begin in June. The Georgia factory that will build the robotaxi version is still under construction. More importantly, Rivian has never tested and deployed a self-driving system designed for robotaxis. The Rivian Autonomy Platform, which debuted in 2024 on second-generation R1 vehicles, handles highway hands-free driving and a point-to-point navigation feature expected later this year. It is not a robotaxi system. It is not close to one. The gap between a highway driver-assist feature and a commercial autonomous taxi is where most autonomous vehicle programs go to die, quietly or loudly.

Scaringe's response to that history is characteristically sanguine. "If you were to look at the progress in autonomy in the last five years and try to use it as a rough metric or gauge to predict the next five years, you would be wildly wrong," he said at SXSW. That is either a profound insight about exponential progress or the thing every autonomous vehicle founder says before the demo works and the deployment doesn't. Probably both.

The autonomy stack question is also unresolved — and underappreciated. Uber will own and operate the fleet through its app. But robotaxis require a brain that can drive them. The TechCrunch reporting notes that Rivian is developing its own system, with a hardware upgrade launching in a version of the R2 in late 2026 that adds a lidar sensor and an "autonomy computer" capable of processing 5 billion pixels per second. Whether that stack will be ready for commercial robotaxi operations in 2028 — or whether Rivian will partner with a third-party AV provider — hasn't been specified. The press release title says "fully autonomous robotaxis." The fine print leaves room for interpretation.

For context: Uber has now signed more than 25 dedicated robotaxi or autonomous vehicle partnerships globally, including deals with Waymo (operating on Uber's app in Austin and Atlanta), Motional, Baidu, and a stake in U.K. startup Wayve. It recently announced a separate partnership with Lucid and Nuro for robotaxis based on Lucid's Gravity SUV, expected to deploy commercially in San Francisco by the end of this year. Uber's model is platform over hardware — it makes money whether the car is driven by a human or a machine. That model works fine as long as the machines eventually arrive.

Rivian's position is different. It needs to prove it can be something more than a niche EV maker with a loyal but small customer base. The R2, its more affordable SUV starting around $45,000, is supposed to be the volume product that changes the company's trajectory. Turning it into a robotaxi is a way to lock in a large order that improves its manufacturing economics — and gives the company a story about the future that investors might actually believe.

Whether that story survives contact with the actual timeline is another question. The factory isn't done. The car isn't in production. The autonomous stack doesn't exist yet. And the deployment target — two cities in 2028, 25 by 2031 — is exactly the kind of number that looks reasonable on a press release and turns into a footnote two years later.