The partnership puts Rivian’s in-house chip and full autonomous stack to work as a robotaxi platform, with commercial deployments planned for San Francisco and Miami in 2028.
Uber has been signing robotaxi deals at a pace that can make any single announcement feel routine. But this partnership with Rivian is structurally different from the rest of the pile. Unlike the company’s deals with Waymo, Avride, or Zoox, where Uber is essentially a distribution platform for someone else’s autonomous vehicle stack, this one bets on a car maker that has built its own silicon, its own autonomy software, and its own manufacturing, end to end.
Rivian (NASDAQ: RIVN) and Uber (NYSE: UBER) announced on Thursday that Uber will invest up to $1.25 billion in Rivian through 2031, contingent on Rivian hitting a series of autonomous performance milestones by specific dates.
An initial $300 million has been committed following signing, subject to regulatory approval. In exchange, Uber, or its fleet partners, will purchase 10,000 fully autonomous Rivian R2 robotaxis in the first phase of deployment, with an option to negotiate the purchase of up to 40,000 more beginning in 2030. The total 50,000 figure is a ceiling, not a guarantee.
Commercial service is planned to begin in San Francisco and Miami in 2028 and expand to 25 cities across the US, Canada, and Europe by 2031. The vehicles will be available exclusively through the Uber platform. All of these timelines are forward-looking and milestone-dependent, and Rivian’s own securities disclosures warn that actual results may differ materially from the companies’ projections.
The deal’s logic rests heavily on what Rivian unveiled at its inaugural Autonomy & AI Day in Palo Alto in December 2025. There, the company laid out a full-stack autonomous driving architecture built around RAP1, its first in-house processor: a custom 5nm chip capable of 1,600 sparse TOPS of AI compute, fabricated by TSMC.
Two RAP1s power the company’s Gen 3 Autonomy Compute Module (ACM3), which can process 5 billion pixels of sensor data per second. The module uses RivLink, a proprietary low-latency interconnect, to allow chips to be chained together for greater compute headroom.
Rivian also developed its own AI compiler and platform software to run on top of the chip, a degree of vertical integration that puts it alongside Tesla as one of the only consumer EV makers designing proprietary silicon specifically for autonomy.
The Gen 3 platform, comprising 11 cameras (65 megapixels total), five radars, and one LiDAR sensor, is currently undergoing validation and is expected to ship on R2 models from late 2026, according to coverage of Autonomy & AI Day by WardsAuto, Edmunds, and Electrek.
Rivian confirmed at the time that the initial R2 production run, expected earlier in 2026, would launch without the Gen 3 hardware. The robotaxi programme announced Thursday is built on the Gen 3 platform, meaning commercial deployments are downstream of that hardware validation completing successfully.
Uber CEO Dara Khosrowshahi pointed specifically to Rivian’s vertical integration as the basis for the company’s conviction. “We’re big believers in Rivian’s approach, designing the vehicle, compute platform, and software stack together, while maintaining end-to-end control of scaled manufacturing and supply in the US,” he said in the announcement.
The framing is notable: unlike Uber’s deal with Lucid and Nuro, where Nuro provides the autonomous driving software and Lucid provides the vehicle, Rivian is building and owns the full stack. Uber is licensing access to it, not integrating two separate companies’ technologies into a single car.
For Rivian, the deal provides both capital and a high-stakes use case for a platform it has spent years building from scratch.
RJ Scaringe, Rivian’s founder and CEO, described the partnership as accelerating the company’s path to Level 4 autonomy. “The scale of Rivian’s growing data flywheel coupled with RAP1, our state of the art in-house inference platform, and our multi-modal perception platform make us incredibly excited for the rapid advancement of Rivian autonomy over the next couple of years,” he said.
Rivian ended Q4 2025 with $6.59 billion in total liquidity, including nearly $6.1 billion in cash and equivalents, providing a runway to absorb the development costs the milestone-based investment structure implies.
The timing sits within a broader burst of Uber robotaxi activity. Over the past six months, Uber has announced partnerships with Zoox (Las Vegas, this summer; Los Angeles, 2027), Wayve and Nissan (Tokyo, late 2026), NVIDIA and Stellantis (28 cities by 2028), and extended its existing arrangements with Waymo, Avride, WeRide, and Lucid/Nuro.
The company has made clear it intends to operate as a multi-vendor platform rather than bet the autonomous mobility transition on any single supplier. The Rivian deal fits that pattern, but with Uber taking a meaningful equity stake, up to $1.25 billion, rather than simply listing another company’s cars in its app.
What remains unresolved is whether Rivian’s autonomous software stack will actually reach Level 4 on the timeline both companies are now publicly committed to. The company’s December 2025 roadmap laid out a progression from point-to-point autonomous driving in early 2026, through eyes-off capability, to personal Level 4, but did not specify firm dates for the later milestones.
