Independent analysis of Tesla robotaxi safety data
This tracker monitors Tesla's robotaxi safety performance in Austin, Texas — the only location where Tesla operates fully unsupervised (Level 4) autonomous vehicles. We calculate miles per incident (MPI) using NHTSA incident reports and fleet mileage data.
Incident data: NHTSA Standing General Order 2021-01 reports, which require manufacturers to report crashes within 1-10 days.
Fleet data: robotaxitracker.com for daily vehicle counts, plus Tesla earnings disclosures.
Human baselines: NHTSA CRSS 2023 and Swiss Re/Waymo study.
We calculate MPI by dividing estimated fleet miles (fleet size x 115 mi/vehicle/day) by the number of NHTSA-reported incidents. The 115 mi/day figure comes from Tesla's Q3 2025 disclosure of 250,000 cumulative miles.
We fit an exponential trend model to track safety improvement over time, reporting R² fit quality and doubling time estimates.
Reporting lag: NHTSA reports can be delayed 1-10+ days, so recent "streak" numbers are provisional.
Fault unknown: SGO reports don't specify fault; some incidents may be other drivers' fault.
Safety monitors: Some Austin vehicles still have monitors who may intervene, potentially inflating safety metrics.
Independent analysis by Kangning Huang
This robotaxi safety tracker was created to provide transparent, data-driven insights into Tesla Cybercab safety performance. It is not affiliated with Tesla, Inc. or any other company.
The project is open source on GitHub. Contributions, bug reports, and data corrections are welcome.