Uber plans to turn its driver network into an AV sensor grid

Uber plans to eventually outfit its human drivers’ cars with sensors to collect real-world data for autonomous vehicle (AV) companies. Praveen Neppalli Naga, Uber’s chief technology officer, revealed the plan in an interview at TechCrunch’s StrictlyVC event in San Francisco on Thursday night. If executed, the strategy would transform Uber’s network of millions of drivers globally into a distributed sensor grid, potentially solving the data collection challenges that currently limit AV development.

The long-term plan builds on a program called AV Labs, which Uber announced in late January. While AV Labs currently uses a dedicated fleet of sensor-equipped cars separate from its driver network, Uber’s ultimate goal is to scale this data collection to its broader driver base. However, Naga noted that the company must first understand how the sensor kits work and navigate regulatory clarity across different states regarding what sensors and data sharing mean.

To support this ecosystem, Uber is building an “AV cloud,” which will serve as a library of labeled sensor data for partner companies to query and use to train their models. Uber currently has partnerships with 25 AV companies, including Wayve, which operates in London. Through this system, partners can also run their trained models in “shadow mode”—a testing method where models run against real-world trips without controlling the vehicle—to simulate how an autonomous vehicle would have performed on real Uber trips.

This data-centric approach represents a strategic shift for Uber, which previously abandoned its own ambitions to build self-driving cars. Instead of competing directly with AV developers like Waymo, Uber is positioning itself as the infrastructure layer for the industry. Naga asserted that the primary limiting factor for AV development is no longer the technology itself. “The bottleneck is data,” Naga said, explaining that AV companies often lack the capital to deploy fleets just to collect diverse real-world scenarios. He added that Uber’s goal is not to make money out of this data, but rather to democratize it for its partners.

Why it matters

Uber is positioning itself as the data layer for the autonomous vehicle industry, potentially solving the data bottleneck that currently limits AV development by leveraging its driver network.