Uber says it isn't building a super-app. Its AV data fleet says otherwise.

Uber chief product officer Sachin Kansal told TechCrunch last week that the company isn’t chasing a super-app. Six months ago, Uber launched a sensor-equipped vehicle fleet, separate from its driver network, designed to gather the real-world driving data that its autonomous-vehicle partners need to train their models.

That fleet is called AV Labs, and it puts a count on the data-collection play: Uber has lined up roughly 20 AV partners, including Waymo, Waabi, and Lucid. The hitch is that some of those partners are also Uber’s competitors in the ride market, with Waymo the most visible overlap. Uber holds equity in several of them, according to TechCrunch’s reporting on the interview. The strategy the public line is deflecting from is the data layer.

The public line is that Uber is a focused mobility-and-travel platform, anchored on the trip. Kansal’s number for the spine is 1.5 billion: that many Uber trips a year happen outside the rider’s home city. Hotels through the Expedia partnership, boat rentals in Europe, a “shop for me” concierge, and the rest of the 2026 expansion slate are presented as compounds on that travel stack. Driver-facing additions like debit cards and a data-labeling side hustle are pitched as financial-services rail for the people who keep the network running. Surface area looks broad. The trip is the anchor.

After Uber’s 2018 pedestrian fatality in Tempe, Arizona, the company shut down its own robotaxi program and sold ATG, its self-driving unit, to Aurora in 2020. AV Labs is a deliberate return to autonomy, but as a data-and-platform play, not as a vehicle operator. The fleet sits between Uber’s mission and its AV partners’ missions: the cars log the road so that partner models can learn from it, and the data stays inside Uber’s walls.

The capital ties tell the same story. Uber owns roughly 11% of Lucid after a $500 million robotaxi investment in April 2026, a stake that Uber and Lucid expanded alongside new PIF funding. Uber joined the Waabi round that valued the Toronto-based AV company at $1 billion in January 2026. The pattern is consistent: small equity checks, larger platform commitments, and access to the data fleet that the partners will use to train against.

A reader can use that as a model for the next Uber announcement. Every new product falls into one of three buckets: a travel-stack compound on the 1.5 billion out-of-home trips, a driver-facing financial-services rail, or an AV data-asset move. The Kansal interview adds color to all three. It does not retire the contradiction.

The contradiction is also not fully resolved in the public record. The 20-partner count for AV Labs is a January 2026 figure that the Kansal interview did not refresh. The data-fleet program could still turn out to be a customer-acquisition cost or a goodwill program for the AV partnerships, rather than a margin lever. The claim that Uber “holds equity in several AV partners” is TechCrunch’s characterization of the Waymo relationship, not an enumerated capital table. The public line and the data fleet can sit in the same story as long as the data is the platform and the platform is the customer.

Uber’s own framing of the autonomous future reads more aspirational than operational. The next AV-partner renegotiation is the first real test of whether the company can keep the data layer and the ride layer from colliding on the same street.