Tesla to Build Its Own AI Maps to Sharpen FSD Navigation

AUSTIN, Texas — Tesla is taking control of one of the last pieces of its self-driving stack it does not yet own: the map. On the company's third-quarter earnings call, Head of AI Ashok Elluswamy said Tesla is building an entirely new, AI-based mapping and navigation system to replace the third-party data it currently licenses from Google.

Owning the Whole Stack

"You're totally right that we shouldn't be depending on external people's maps," Elluswamy said in response to a question about recurring navigation issues, including speed-limit recognition, parking and routing. "We are building an entirely new architecture with essentially AI-based mapping and navigation."

The move extends a philosophy that has defined Tesla's approach to Full Self-Driving: vertical integration. By generating its own maps from the sensor data its vehicles already collect, Tesla can iterate as fast as it ships software, rather than waiting on an outside provider to refresh the underlying data.

A Fleet That Maps Itself

Elluswamy described a system that improves automatically as cars drive. "The collective intelligence of all the cars is going to flow into making the maps and navigation super cool," he said, adding that the mapping will be integrated with voice so drivers can "talk to it, ask for different routes."

That fleet-learning loop is the same advantage that powered Tesla's camera-based driving system, and the company expects it to resolve the routing and parking complaints that a routing feature in the Summer 2026 Update only partially addressed. The timing dovetails with Tesla's broader autonomy push, including its fast-growing robotaxi operations.

Arriving With v15

Elluswamy said the new mapping should be available around the time Tesla's v15 software begins deploying to customers, and that over the following months it would progressively take over navigation, parking and destination handling. Tesla confirmed the plan in comments reported by Teslarati.

Replacing a core dependency with an in-house, self-improving system is classic Tesla — and if the AI maps deliver on Elluswamy's description, drivers could soon see navigation that sharpens with every mile the fleet travels, closing one of the few remaining gaps between Tesla's software and a truly seamless autonomous experience.