Tesla Says It Will Construct Own Mapping Data for FSD
The automaker's leadership acknowledged during a recent earnings call that proprietary navigation data is required to address specific shortcomings in its autonomous driving technology.
By Alina Carter ·
Tesla has announced plans to develop its own proprietary mapping infrastructure to correct persistent issues within its Full Self-Driving suite, specifically concerning the reading of speed limits, parking protocols, and route determination. The company's Vice President of AI, Ashok Elluswamy, addressed the inquiry during the shareholder Q&A portion of the Q3 2026 earnings call. Elluswamy stated that the company 'shouldn't be depending on external people's maps' and confirmed that Tesla is building its own maps using a new AI-based mapping and navigation architecture.
The executive outlined that this technology will begin handling the initiation and conclusion of trips with FSD v15. Over the subsequent months, the system is intended to assume control over navigation, parking, and destination handling. This development marks a significant pivot for a company whose CEO has spent years arguing that self-driving systems relying on maps are brittle. It also indicates that Tesla's fleet is now performing the same function as Waymo's vehicles: mapping roads so the self-driving system can navigate areas it cannot reliably determine independently.
The company's history regarding cartography has been inconsistent. In 2015, when launching Autopilot, Tesla told Fortune it had begun building high-precision maps with '100 times the level of granularity' of standard navigation. However, at Autonomy Day in 2019, Elon Musk stated that Tesla had attempted high-precision lane maps and realized it was 'a huge mistake.' He described reliance on such data as a 'really bad idea' that makes a self-driving system 'extremely brittle,' arguing instead for a vision-based approach.
The specific problems Tesla is attempting to resolve involve understanding the environment, an area where FSD reportedly fails, particularly regarding speed limit signs. Recent data highlights these difficulties. Last month, a Belgian road safety group found that FSD operated at unsafe speeds in 55 percent of 30 km/h zones in Brussels, with the vehicle's display often showing an incorrect limit. Additionally, the National Highway Traffic Safety Administration is considering an investigation into reports of Tesla Robotaxis speeding in active school zones.
If a vision system cannot reliably read and apply basic road information, and the solution requires a better map, then vision alone may not be sufficient for the task. This shift brings Tesla closer to Waymo's operational model than it would likely prefer to admit. Waymo maps cities by driving sensor-equipped vehicles before launching service and maintains those maps as its fleet drives. Tesla is now describing a similar process where its cars collect data to feed a Tesla-built map that assists all vehicles.
While Tesla has always used navigation maps, building its own rather than licensing them from third parties represents a change in strategy regarding the source of that data. A map created from millions of Tesla vehicles could provide a significant advantage if it resolves issues regarding speed limits. However, this approach contradicts previous messaging that dismissed detailed maps and lidar as unnecessary for autonomous driving.
Waymo constructed its maps and has been operating a driverless service for years. Tesla is constructing its maps now and has scheduled unsupervised FSD for consumer cars in Q2 2027, initially limited to its Robotaxi fleet. The company's current operations are confined to geofenced areas using maps, a method Musk previously stated does not constitute a 'true self-driving system.'
