Tesla has started offering paid robotaxi rides with no one in the driver’s seat in Austin, Texas. The service uses a mix of supervised and unsupervised Full Self-Driving (FSD) vehicles, and the unsupervised group is still small.
Company executives describe this as a step‑by‑step shift rather than a full launch. Ashok Elluswamy, Tesla’s vice president of AI software, said the company is “starting with a few unsupervised vehicles mixed in with the broader robotaxi fleet with safety monitors, and the ratio will increase over time.”
In the new setup, some trips run with no human supervisor inside the robotaxi. Riders can sit in the front passenger seat while FSD controls the steering, acceleration and braking.
Elon Musk announced the change on X and wrote that Tesla had “just started Tesla Robotaxi drives in Austin with no safety monitor in the car” and thanked the AI team. Markets reacted fast, and trading data indicate Tesla’s share price climbed more than 4% after the news.
Earlier in 2025, Tesla began with supervised robotaxi trips in Austin, where a safety monitor sat in the front passenger seat and could take over at any moment. Later, the company quietly tested empty vehicles before opening unsupervised rides to paying customers.
Video: Tesla unsupervised driving
In passenger seat now pic.twitter.com/rjfkos2ZCe
— TSLA99T (@Tsla99T) January 22, 2026
Musk’s targets and Cybercab plans
Musk has repeated that Tesla aims for wide robotaxi coverage in the United States by late 2026. In recent comments, he said he expects the service to become “very, very widespread” and has described high‑level autonomy as close to solved from a technical point of view.
For now, the driverless rollout remains limited to a small service area. Tesla is also working on the Cybercab, a purpose‑built robotaxi with no steering wheel or pedals, planned for production at the Texas factory starting in 2026, with higher volumes targeted for 2027.
Tesla wants Cybercab operating costs near 20 cents per mile, a figure that would undercut many ride‑hailing fares if achieved.
Different bets on technology
Tesla stands apart from many rivals on its technical approach. The company now relies on a camera‑only sensor suite and large neural networks, dropping lidar and radar from its current consumer and robotaxi hardware.
Tesla says it trains FSD on billions of miles collected from cars on public roads, with company materials putting total FSD‑assisted mileage in the billions by 2026.
Competitors such as Waymo and Cruise continue to use multiple sensor types and often limit service to tightly mapped zones.
The Austin launch gives Tesla real‑world evidence on how unsupervised FSD performs with paying riders.
For now, investors are tracking the data, safety agencies are watching the streets, and Tesla is pressing its case that software can take over more of the driving task.
You may also like:
- Musk says Tesla’s HW4 cars are ready for unsupervised FSD »
- Elon Musk sets new data target for Tesla unsupervised self‑driving »
- Musk says unsupervised FSD is “pretty much solved” »
We’ve launched an official Tesla Owners Online Forum. Join early and claim your "Founding Fuel" badge.
