Tesla's Shift: Unsupervised Driving Limited to Robotaxi Zones

Tesla has officially set a new target of the second quarter of 2027 for unsupervised Full Self-Driving in vehicles owned by customers, marking the latest adjustment to a promise the company has made for nearly a decade. However, this timeline comes with a significant structural change that redefines what the feature actually entails for the average driver.

On the company’s third-quarter 2026 earnings call, AI chief Ashok Elluswamy clarified that the upcoming software update will not grant owners the ability to drive their cars anywhere they wish without supervision. Instead, the plan involves integrating customer-owned vehicles into Tesla’s existing managed Robotaxi fleet, meaning the cars will operate under Tesla’s control rather than as independent autonomous units.

This distinction addresses a primary concern raised by retail shareholders during the earnings preview, where questions about the arrival of unsupervised FSD in personal cars were backed by approximately 1.1 million shares. Elluswamy’s response indicated that the first step is not a universal software unlock, but rather a mechanism for owners to contribute their vehicles to the Robotaxi network.

Consequently, an owner’s car will only be capable of unsupervised driving within the specific geographic boundaries where Tesla’s Robotaxi service is already active. Currently, this service operates in six US metropolitan areas, but the actual service areas are geofenced to small, manageable slices of those cities. For example, the Miami zone largely covers West Miami and a limited strip toward Doral and Sweetwater, while Orlando and Tampa launched with similarly restricted perimeters in July.

Within these designated zones, the vehicles follow low-speed routes, avoid highways, and steer clear of complex urban environments. A notable exception is the Bay Area, where the service is not yet unsupervised due to California regulations requiring a safety driver in every vehicle, a requirement Tesla maintains because it has not yet released the necessary data to the California DMV for a level beyond two autonomous systems.

The scale of the current operation remains modest relative to previous executive statements. Tesla reported that its Robotaxi fleet reached nearly two million cumulative unsupervised miles by the end of the third quarter, a month after the Cybercab joined the Austin fleet. Third-party tracking data from earlier this spring suggested the number of fully driverless Teslas on the road was in the dozens, a figure far below the hundreds of thousands previously discussed by CEO Elon Musk.

Competitors such as Waymo had already established a presence in both Orlando and Tampa before Tesla’s entry, with Waymo reporting over three million miles of operation. The gap between the original marketing pitch and the current product reality is widening, as previous promises included a car that would drive itself anywhere via a simple update, whereas the current plan requires the car to function as part of a controlled fleet in specific zones.

Elluswamy identified safety as the primary bottleneck for expansion, explaining that Tesla limits operations to lower-speed roads and simpler environments until further progress is made. He estimated that these safety issues could be resolved by the following year, though he acknowledged uncertainty in that projection. For owners of hardware version three vehicles, the situation remains complicated, as they have been informed their computers cannot run unsupervised FSD without a clear upgrade path.

Ultimately, the new timeline and restrictions represent a shift in the goalposts for FSD buyers. Rather than a vehicle that drives anywhere, the product is now defined by its ability to join Tesla’s fleet in specific zones under Tesla’s terms. This approach eliminates highway trips and cross-town commutes unless the user’s daily life is entirely contained within one of the limited service areas.