Uber wants to turn millions of drivers into a grid of sensors for the self-driving company.

Uber has long-term ambitions beyond transporting passengers. The company ultimately wants to equip human drivers’ vehicles with sensors to absorb real-world data for autonomous vehicle (AV) companies and other companies training AI models in real-world scenarios.

Uber Chief Technology Officer Praveen Neppalli Naga revealed the plan in an interview at TechCrunch’s StrictlyVC event in San Francisco on Thursday night, describing it as a natural extension of an earlier program called AV Labs that the company announced in late January.

“That’s ultimately the direction we want to go,” Naga said of equipping vehicles with human drivers. “But first, you need to understand sensor kits and how they work. There are some regulations. Every state needs to be (clear) what a sensor means and what sharing means.”

Currently, AV Labs relies on a small, dedicated fleet of sensor-equipped vehicles that Uber operates independently of its driver network. But the ambitions are clearly much greater. Uber has millions of drivers around the world, and if even just a few of those cars can be converted into a rolling data collection platform, the scale of what Uber can offer the AV industry will dwarf what individual AV companies can assemble on their own.

Naga said the insight driving the program was that the limiting factor in AV development was no longer the underlying technology. “The bottleneck is data,” he said. “(Companies like Waymo) have to go around and collect data and collect different scenarios. In San Francisco, they might say, ‘I need some data from this time of day so I can train a model on this school intersection.’ The problem for these companies is access to that data, because they don’t have the capital to deploy the cars and collect all that information.”

Becoming the data layer of the entire AV ecosystem is a very smart play. Especially considering that a few years ago Uber abandoned its own ambitions to build self-driving cars (a move that co-founder Travis Kalanick publicly lamented as a big mistake). In fact, many industry observers have wondered whether without its own self-driving cars, Uber could one day become irrelevant as AVs increasingly proliferate around the world.

The company is currently partnering with 25 AV companies, including Wayve, which operates out of London, to build what Naga describes as an “AV cloud”. This is a library of labeled sensor data that partner companies can query and use to train models. Partners, whom Uber plans to invest more actively in directly, can use the system to run trained models in “shadow mode” on real Uber trips, simulating how AVs will perform without actually putting them on the road.

“Our goal is not to make money from this data,” Naga said. “We want democracy.”

Given the clear commercial value of what Uber is building, this positioning may not last long. The company has already made equity investments in a number of AV players, and its ability to provide proprietary training data at scale could give it significant leverage in a sector that currently relies on Uber’s ride-hailing market to reach customers.