IT之家 AISource publication

Tesla confirmed that it is developing its own AI map, which will be rolled out gradually with FSD v15

IT Home On October 25, during the latest Tesla earnings call, Tesla Vice President Ashok Elluswamy announced that Tesla is developing a custom AI-based map navigation system. This new map solution will be released with FSD v15, integrating data collected by the fleet and personalized route planning, while also supporting FSD voice control.

IT Home On October 25, during the latest Tesla earnings call, Tesla Vice President Ashok Elluswamy announced that Tesla is developing a custom AI-based map navigation system. This new map solution will be released with FSD v15, integrating data collected by the fleet and personalized route planning, while also supporting FSD voice control.

According to IT Home, Tesla’s current navigation uses third-party map content combined with its own developed software. When the “preferred route” feature was introduced earlier, reports indicated that Google provided base maps, street data, satellite images, points of interest, and traffic information through API. Tesla develops its own navigation app and calculates routes, so it can incorporate ultra-fast charging stations and user preferences into route decisions.

At the earnings meeting, a shareholder asked about the issues faced by FSD in speed limit recognition, parking, and route selection. Eluvsmami agreed: “You are completely right; we shouldn’t rely on external manufacturers’ maps at all.”

He later confirmed that Tesla is developing its own AI map system to address a series of issues exposed by FSD, including speed limit recognition, parking, and route planning.

Eluvsmi said this is a completely new architecture, not just another version of a map. Essentially, this solution is a “AI-based map and navigation system”. In other words, Tesla’s large fleet will automatically collect information about the streets, including data on traffic speeds, parking signs, traffic lights, and parking lots.

The new map architecture will unlock a lot of new capabilities. Tesla can overlay rich information onto the map, not only solving the pain points of FSD such as routes and speed limits, but also making driving smoother and safer: marking potholes, entry and exit points with blocked views, dangerous intersections, and so on.

It is highly likely to be used for optimizing automatic parking, allowing users to customize their parking preferences in the parking lot: prioritize the nearest empty spaces, stay away from other vehicles, and choose accessible parking spots. The fleet will continuously collect relevant environmental data.

Eluvsmi also mentioned that with this map, users can customize routes using voice, which aligns with the Grok capability that will be implemented in autumn this year to control FSD.

Tesla has introduced the “Preferred Route” feature in the 2026 summer update, allowing users to prioritize routes they have previously taken when planning a journey. In the same update, automatic navigation capabilities were added, enabling automatic navigation to other frequently visited locations, in addition to home, work, and calendar trips.

And the preferred route is an iteration based on the alternative route feature that was launched in 2022. The old feature offered at most three alternative routes.

Elukswami expects that this new map will be used first in FSD v15 for the departure and arrival stages of trips, and then over the next few months it will gradually cover the remaining navigation functions, including parking and voice control.

Original source

IT之家 AI

Content notes

Original publication and rights belong to the source.

Machine translation · Refer to the original