Uber is building its own fleet of sensor-equipped vehicles to gather massive amounts of autonomous driving data, a move that gives the ride-hailing giant both leverage and optionality in its increasingly complicated relationship with robotaxi partners like Waymo. The company's six-month-old AV Labs unit represents a strategic shift: rather than relying entirely on external autonomous vehicle providers, Uber is positioning itself as a data partner that can offer insights no other ride-hailing platform can match.

What Is Uber's AV Labs, and Why Does It Matter?

AV Labs is a separate business unit within Uber that operates hundreds of sensor-equipped vehicles deployed through fleet partners. The goal is straightforward but ambitious: collect millions of miles worth of driving data to help solve what the autonomous vehicle industry calls the "long-tail problem." This refers to the rare, edge-case scenarios that self-driving systems need to handle but rarely encounter during normal testing.

Uber frames this initiative publicly as a way to strengthen relationships with its autonomous vehicle partners, several of which it also holds equity stakes in. But the company's Chief Product Officer Sachin Kansal acknowledged the deeper strategic value. "We are going to be equipping hundreds of cars with sensors, deployed through our fleet partners, and through that we'll be collecting millions of miles worth of driving data," Kansal explained. "That really helps with the long-tail problem, you want to see all the edge cases, not just the P95, P99 level".

"We are not in the race to be an L4 autonomy provider, what we are focusing on is laying down the race tracks so we can work with multiple players," said Sachin Kansal, Chief Product Officer at Uber.

Sachin Kansal, Chief Product Officer at Uber

How Does Uber's Data Strategy Create Competitive Advantage?

The data layer gives Uber something its autonomous vehicle partners don't easily have access to: operational expertise from managing 10 million drivers and handling the messy realities of ride-hailing at scale. Kansal noted that Uber processes 25 million lost items annually, manages millions of pickups and drop-offs daily, and understands the logistical challenges that autonomous fleets will face in the real world.

This operational knowledge, combined with raw driving data, creates a valuable asset. Uber can offer autonomy partners insights into how to actually run a robotaxi service, not just how to make a self-driving car work in a lab. The company is essentially positioning itself as a bridge between autonomous vehicle technology and the messy reality of commercial ride-hailing operations.

Why Is Uber Competing With and Partnering With Waymo Simultaneously?

Uber's relationship with Waymo exemplifies the tension. The two companies partner in some cities, yet compete directly in others. Uber recently wound down its robotaxi pilot with Waymo in Phoenix, where the two had operated about a dozen autonomous vehicles together. However, Uber has scaled up Waymo partnerships in Austin and Atlanta, where it now operates hundreds of autonomous vehicles.

Kansal explained Uber's hybrid approach: the company believes that mixing human drivers and autonomous vehicles in the same city allows it to balance supply and demand more effectively. This strategy gives Uber flexibility that pure autonomous vehicle operators don't have. By maintaining relationships with multiple autonomy providers while building its own data operation, Uber avoids becoming dependent on any single technology partner.

Ways Uber Is Expanding Beyond Ride-Hailing and Delivery

• Travel Services: Uber added hotel bookings through a partnership with Expedia, recognizing that 1.5 billion trips annually occur outside users' home cities, making travel a major use case for the platform.
• Financial Products for Drivers: Uber Pro card allows drivers and couriers to use it as a debit card and transfer earnings directly, with the company experimenting with similar products for merchants in certain regions.
• Membership Cross-Selling: Uber One membership has reached 51 million members and accounts for roughly half of all bookings, with data showing that members increase usage across multiple service lines after joining.

Uber's expansion strategy reflects a deliberate choice not to become an "everything app" in the vein of Asian super-apps. Instead, the company is selectively adding services that complement its core ride-hailing and delivery businesses. Kansal stated that Uber is "not trying to be everything to everyone," suggesting the company will remain disciplined about which new services it pursues.

Kansal

The AV Labs initiative fits this broader strategy. By controlling the data layer, Uber maintains optionality. If one autonomous vehicle partner falters, Uber has the data and operational expertise to work with alternatives. If Waymo or another partner becomes too dominant, Uber's own data operation provides leverage in negotiations. This is the kind of strategic hedging that separates ride-hailing platforms from pure technology providers in the autonomous vehicle race.