Uber’s CTO, Praveen Neppalli Naga, dropped a bombshell at TechCrunch’s StrictlyVC event: the company wants to eventually outfit its millions of human drivers’ cars with sensors to vacuum up real-world data for autonomous vehicle companies. This isn’t just a side hustle. It’s a strategic pivot. After abandoning its own self-driving ambitions years ago, a move co-founder Travis Kalanick has called a colossal mistake, Uber is now aiming to become the indispensable data broker for the entire AV industry. The program, called AV Labs, currently runs a small dedicated fleet, but the vision is to scale it across the globe. Naga claims the goal is to democratize data, insisting Uber isn’t looking to make money directly from it. But given the company’s equity stakes in AV players and the immense commercial value of proprietary training data, that’s a hard sell. This is a play for control, dressed in altruistic clothing.
The Data Fiefdom Uber is Building
Naga stated bluntly that the bottleneck for AV development is no longer technology, but data. Companies like Waymo struggle to gather enough diverse scenarios, from specific school intersections to tricky weather conditions. Uber’s solution is to build an ‘AV cloud’, a library of labeled sensor data that its 25 partner AV companies can query and use to train their models. This includes a ‘shadow mode’ where partners can run their trained models against real Uber trips without putting vehicles on the road. Essentially, Uber is creating a walled garden of training data that the AV industry will have to pay or partner to access. It’s a brilliant, if cynical, strategy to make itself the gatekeeper of the physical world’s digital twin, leveraging drivers who are already being squeezed by the platform.
The Creep Factor of a Million Rolling Surveillance Rigs
Let’s not ignore the elephant in the car. Outfitting millions of vehicles with lidar, cameras, and radar isn’t just about collecting traffic patterns. It’s a massive, decentralized surveillance network. Uber insists it will navigate state regulations about sensor data and sharing, but the scale of this operation is unprecedented. While Naga frames it as a data-sharing ecosystem for AV safety, the potential for mission creep is staggering. Who else might want access to this firehose of street-level data? Insurance companies? Urban planners? Law enforcement? Uber’s claim that it wants to ‘democratize’ this resource rings hollow when the company itself controls the pipes. For a company with a controversial history on privacy and driver treatment, this feels less like a public good and more like a data grab with a PR gloss.
Source: Techcrunch
