Uber launches robotaxi support project for more than 20 AV partners

Uber is rolling out a new effort to support third-party robotaxi companies, pitching what it calls "a comprehensive suite of unique services and capabilities that are already helping partners to build and successfully commercialize autonomous vehicles in multiple markets around the world." The initiative appears under two names in public materials - Uber Autonomous Solutions and Uber AV Labs - and will offer infrastructure, user experience and fleet-management tools rather than building vehicles itself.

The project will be led by Sarfraz Maredia, identified as Global Head of Autonomous Mobility and Delivery at Uber, and is already linked to a concrete passenger pilot in Dallas with Avride. Dmitry Polishchuk, Avride’s chief executive, said the company is expanding from autonomous delivery into passenger mobility and is "excited to begin introducing [robotaxis] in Dallas, with our partners at Uber." Riders in Dallas are being asked to opt in under Ride Preferences in the Uber app to boost their chances of matches.

Uber’s AV Labs component is collecting driving data with a prototype fleet that currently includes one car and plans to add more. A leader quoted as Naga described the operation as intentionally scrappy: "We don’t know if the sensor kit will fall off, but that’s the scrappiness we have," and added that it could take time before the program scales, "I think it will take a while for us to say, deploy 100 cars to the road to start collecting data. But the prototype is there."

The collected sensor streams will not be passed to partners as raw feeds. Instead, AV Labs will process and refine the information into a "semantic understanding" layer intended to feed driving software and improve real-time path planning. Uber has already engaged with more than 20 autonomous vehicle partners and has discussed the service with companies including Waymo, Waabi and Lucid Motors, though several of those relationships remain exploratory and contracts have not all been finalized.

The program arrives as the AV industry shifts from rules-based systems toward reinforcement learning approaches that depend on large quantities of real-world driving data. Nvidia, in a press release dated October 28, 2025, framed the move as part of a growing Level 4 ecosystem and said it is supporting global robotaxi expansion in collaboration with Uber and others. Nvidia also announced a Halos Certified Program it described as the industry's first system to evaluate physical AI safety for autonomous vehicles and robotics.

Uber’s corporate materials position the new work as complementary to human drivers. Sarfraz Maredia said the company is "proving how AVs and drivers can work side by side to make transportation more convenient, sustainable, and affordable for people everywhere." Uber’s rider-facing materials stress safety, noting that "any AV on the Uber network must meet all our Safety Guidelines" and that riders will have access to human support through the app.

The effort recalls Uber’s earlier experiment in building robotaxis, which the company stopped after a 2018 test-vehicle fatality and later divested in 2020. Today’s strategy focuses on selling operational capabilities to vehicle makers and robotaxi operators rather than competing as a vehicle manufacturer.

Key questions remain about the new work’s scope. Public materials use multiple names for the initiative and leave unclear whether AV Labs is a distinct data arm of a broader Uber Autonomous Solutions product suite. The size and timeline for scaling beyond the one-car prototype, the exact format of the semantic data that partners will receive, and the commercial terms of partner agreements have not been fully disclosed. As Uber expands pilots such as the Dallas Avride program, those details will determine whether the company can convert operational scale and rider reach into a practical advantage for the next generation of robotaxis.