Waabi expands to Uber robotaxis in US$1bn funding round

Waabi has raised US$1bn in combined funding and struck a partnership with Uber to deploy robotaxis on the ride-hailing platform, marking the Canadian firm’s first expansion beyond trucking. The funding consists of a US$750m Series C round co-led by Khosla Ventures and G2 Venture Partners plus around US$250m in milestone-based capital from Uber to support the deployment of 25,000 or more Waabi-powered robotaxis.

Waabi robotaxis will subsequently be rolled out exclusively on the Uber platform. Neither company has yet disclosed timelines for when this may begin. The development is only the latest in a long series of moves by Uber to partner with—and often invest in—autonomous driving companies for platform-exclusive robotaxi services. 

For North American deployments it counts Waymo, Lucid, Nuro, Avride and Volkswagen among its partners; for Europe it has tied up with Wayve, WeRide, Momenta and Baidu’s Apollo Go; in the Middle East it collaborates with WeRide, Pony.ai, and Apollo Go. In the case of Waabi, no specific regions have yet been confirmed, but North America is a likely candidate given the firm’s Canadian roots.

The latest investment round lands Waabi at a valuation of approximately US$3bn, with other investors including Nvidia’s venture capital arm NVentures, Volvo Group Venture Capital, Porsche and BlackRock.

Notably, Waabi Founder and Chief Executive Raquel Urtasun previously served as Chief Scientist at Uber Advanced Technologies Group before its closure in 2020. The unit, which spent considerable resources developing self-driving technology in-house before being deemed too costly by Chief Executive Dara Khosrowshahi, would later see its technology sold to autonomous trucking firm Aurora Innovation. 

In an interview with TechCrunch concurrent to the investment round, Urtasun claimed that Waabi’s “physical AI platform” enables a single technology stack to handle both autonomous trucking and robotaxis using identical underlying models that generalise across vehicle types. This contrasts with competitors like Waymo that previously attempted both verticals before shutting down freight operations. 

The company trains its system using a closed-loop simulator that automatically builds digital twins from data and stress-tests scenarios without human intervention, allowing learning from fewer examples than traditional approaches. “We don’t need the gazillion humans to develop the technology and the large fleets that AV 1.0 needs,” Urtasun told the outlet. “We don’t need the massive data centers, energy consumption, or a gazillion latest chips.”