Tesla Robotaxi Markets: Geofences From 23.6 to 1,098.4 sq. Miles

Tesla’s autonomous ride-hailing program has expanded from a pilot in Austin to a commercial network covering seven major U.S. metropolitan areas. Initial trials used human safety monitors in the driver or passenger seat, but several regions now offer fully driverless service, while one market still requires in-vehicle monitors.

For riders, availability hinges on location. Below is a market-by-market summary of all seven active territories, their geofenced areas, and the types of vehicles in use.

Austin, Texas: 246 sq. miles

Austin is the flagship test bed and the most adaptable market. The geofence spans 246 sq. miles (637 sq. km), covering much of the urban core from North Austin and Tech Ridge through South Austin and Del Valle, with approximate bounds along Highway 183, Highway 71, and State Highway 130.

At launch as a beta in June 2025, service covered just 20 sq. mi. It has since grown to more than 12x its original size through multiple expansions, the latest occurring last month. Austin is also the first city where riders can hail steering wheel-less Cybercabs alongside unsupervised Model Y robotaxis, with Cybercabs offering lower per-mile fares in downtown corridors, East Riverside, and nearby suburban areas.

San Francisco Bay Area, California: 1,098.4 sq. miles

Northern California is the largest market by land area at 1,098.4 sq. miles (2,845 sq. km). Coverage reaches from Richmond, Berkeley, and Oakland across the East Bay, through Fremont and Milpitas into San Jose, and across the peninsula through Palo Alto, San Mateo, Daly City, and San Francisco. Service includes San Francisco International Airport and San Jose International Airport.

Unlike Texas and Florida, operations here are limited to Model Y vehicles with in-vehicle safety monitors while approvals for commercial driverless service advance through state agencies. The California DMV has currently only given Tesla a testing permit that requires a driver even if the vehicle is driving itself.

Dallas, Texas: 31 sq. miles

Dallas operates within 31 sq. miles (80.1 square km), centered on Downtown and extending north through Oak Lawn, Highland Park, University Park, and Bluffview. The boundary sits alongside Dallas Love Field to the west and Lower Greenville to the east.

Rides are fully unsupervised with retrofitted Model Y vehicles, connecting key commercial areas and healthcare destinations such as Parkland Health.

Houston, Texas: 25 sq. miles

Houston’s 25 sq. miles (64.6 square km) service zone sits in the northwest of the city, encompassing Jersey Village, Harvest Bend, Windermere Lakes, and North Houston Gardens. It runs along Highway 290 and extends toward Willowbrook Mall near State Highway 249.

Trips are driverless via Model Y vehicles across residential neighborhoods and nearby retail corridors.

Miami, Florida: 30.8 sq. miles

Miami marks the company’s first Southeast market with a 30.8 sq. mile (79.9 sq. km) corridor. The geofence sits just west of downtown, wraps Miami International Airport, and extends south into Coral Gables, bounded by Highway 836 and Highway 41.

Model Y rides operate without a safety driver, serving airport travelers and adjacent business districts.

Orlando, Florida: 26.2 sq. miles

Orlando’s 26.2 sq. miles (67.9 sq. km) geofence spans eastern neighborhoods just north of Orlando International Airport, including Conway and Azalea Park, with coverage along Semoran Boulevard (State Road 436). State Road 408 forms the northern edge, and State Road 417 marks the eastern boundary. Service runs unsupervised.

Tampa, Florida: 23.6 sq. miles

Tampa covers 23.6 sq. miles (61.2 sq. km) east of Tampa International Airport, including Egypt Lake-Leto, Old Seminole Heights, and Tampa Heights, extending toward the north edge of downtown near Interstate 275 and Highway 92. Riders can request unsupervised Model Y trips.

Ready to Scale With FSD v15

Although these geofences appear compact compared with long-running rivals like Waymo, they serve as focused validation zones ahead of broader rollouts.

Tesla plans to release FSD v15 to the Robotaxi fleet in October. The software introduces a 10-billion-parameter neural network — up from roughly one billion on current branches — and is expected to unlock 24/7 service hours.

The company has said FSD v15 will be the linchpin for scaling Robotaxi to dozens of additional markets. It has been preparing to bring service to Las Vegas, Nevada, and Phoenix, Arizona, with more cities in states that already host Robotaxi operations also under consideration. As commercial miles accumulate, existing geofences are expected to expand and new ones to be added.