Uber's Big Bet on Crowdsourced Road Data
Uber has a new pitch for the self-driving industry: forget expensive robotaxi fleets and closed test tracks. Why not use the millions of human drivers already crisscrossing every city on Earth?
Praveen Neppalli Naga, Uber's chief technology officer, laid out the vision at TechCrunch's StrictlyVC event in San Francisco on Thursday night. The idea is to evolve a program called AV Labs — first announced by the company in late January — into something far more ambitious: a living, breathing sensor grid powered by Uber's existing driver network.
What Is AV Labs?
AV Labs is Uber's framework for partnering with autonomous vehicle companies, giving them access to Uber's platform infrastructure, trip data, and logistics expertise. But according to Naga, that's just the beginning. The next step is turning everyday Uber drivers into passive data collectors for self-driving AI systems.
In practical terms, this could mean equipping vehicles with cameras or other sensors that record road conditions, traffic patterns, pedestrian behaviour, construction zones, and the thousands of edge cases that autonomous vehicles need to learn before they can navigate the world safely. Drivers go about their normal routes — and AV companies get a firehose of real-world training data in return.
Why This Matters for the AV Industry
Building a self-driving car is, at its core, a data problem. Autonomous systems need to encounter millions of scenarios — a child darting into the road, a cyclist running a red light, a pothole hiding under a puddle — before they can handle them reliably. Generating that data through dedicated test fleets is slow and extraordinarily expensive.
Uber's proposition flips the model. Rather than sending out a small fleet of purpose-built sensor cars, the company is offering AV developers instant access to a network already operating at massive global scale. Uber's drivers collectively log an enormous number of trips every single day across hundreds of cities, covering urban cores, suburban sprawl, and everything in between.
For autonomous vehicle startups and established players alike, that kind of geographic and situational diversity is essentially impossible to replicate on their own.
A Natural Extension — or a Privacy Minefield?
Naga framed the move as a natural evolution of what Uber already does — aggregating mobility data — rather than a dramatic new direction. But the concept raises obvious questions about driver consent, data ownership, and passenger privacy. If sensors are recording video or environmental data inside or outside a vehicle during a paid trip, who owns that information, and what rights do riders have?
Uber has not yet publicly detailed the opt-in mechanics for drivers or the data-sharing agreements it would require of AV partners under the expanded AV Labs program.
The Bigger Picture
The announcement is another sign that the lines between traditional rideshare platforms and the autonomous vehicle industry are blurring fast. Rather than competing directly with Waymo or Tesla's robotaxi ambitions, Uber appears to be repositioning itself as the infrastructure layer — the rails that AV companies run on, not just a rival trying to build the trains.
Whether drivers see any financial upside from becoming nodes in a global sensor network remains to be seen.
Source: TechCrunch, StrictlyVC San Francisco, May 2026.
