Uber’s Product Chief on AI Strategy: Robotaxis, AV Labs Data Operation, and Agentic Mobility

Uber Chief Product Officer Sachin Kansal sat down with TechCrunch for an extensive interview covering the company’s strategic bets across multiple frontiers. Uber is rolling out hotel bookings through an Expedia partnership, expanding financial services with the Uber Pro debit card, navigating robotaxi deployments, and launching AI-driven user experiences — though Kansal made clear Uber does not want to become an “everything app.”

In autonomous driving, Uber’s relationship with Waymo has grown increasingly complex. Uber wound down the Waymo pilot in Phoenix while scaling deployments to hundreds of vehicles in Austin and Atlanta. “Waymo is an excellent partner of ours, but in many cities they’re also a competitor,” Kansal said, capturing the delicate dynamic of co-opetition that defines Uber’s autonomous vehicle strategy.

Uber’s AV Labs unit, formed six months ago, is deploying hundreds of sensor-equipped vehicles through fleet partners to collect millions of miles of driving data. Kansal explained this data addresses the “long-tail problem” in autonomous driving — capturing edge cases beyond the 95th or 99th percentile. This data layer gives Uber both leverage and optionality in its relationships with autonomous vehicle partners.

AI is already reaching Uber users. An “earner assistant” helps drivers optimize earnings by analyzing real-time demand data and suggesting higher-opportunity areas. Passengers can use voice to request rides with complex specifications like “I’m looking for a ride to the airport, I have six pieces of luggage, six people.” Kansal described a fully agentic Uber — “plan and book my whole trip” — as on the horizon, though he declined to set a timeline, emphasizing the importance of shipping reliable experiences.

On data monetization, Uber is labeling data for generative AI companies using its earner base for audio collection and annotation. Kansal confirmed this is already a revenue-generating business that the company is “extremely bullish about.” AV Labs’ data-sharing models with partners are still being figured out.

Uber One now has 51 million members, accounting for roughly half of Uber’s bookings. Kansal noted Uber Eats has been independently profitable for several quarters and no longer relies on ride-hailing to stay healthy.

Strategically, Uber is building a three-layer moat: core ride-hailing and delivery network effects, AV Labs’ autonomous driving data assets, and AI-powered experience upgrades. Before the robotaxi era fully arrives, Uber is using data and AI to fortify its platform position.