According to a chart in Tesla’s Q4 2025 earnings report showing cumulative robotaxi miles, the fleet has traveled approximately 500,000 miles as of November 2025. That works out to roughly one crash every 55,000 miles.
For comparison, human drivers in the United States average approximately one police-reported crash every 500,000 miles, according to NHTSA data.
That means Tesla’s robotaxis are crashing at a rate 9 times higher than the average human driver.
That comparison requires there only be a single RoboTaxi though… It’s a faulty comparison.
Tesla’s Robotaxis suck, but you don’t have to go about making questionable statistics to show that.
The statistics make sense. The metric is total num crashes divided by total miles by fleet. It does not matter if one Tesla drove 10 km, or two teslas drove 5 km each.
It does not matter if one Tesla drove 10 km, or two teslas drove 5 km each.
It absolutely does, as they could have an accident with each other. If your statistics were taken, we could simplify down to a single vehicle globally, which can’t have an accident with another vehicle as we’re just removed them from consideration.
No it doesn’t. This effect only comes into effect if there are sufficient teslas on the road that p² is comparable to p, where p is the ratio of teslas to all cars. if 5% of all cars are Teslas, then p is 0.05 and p² is 0.0025, which is a negligible probability. Teslas are much more likely to encounter non-Teslas than Teslas.
That comparison requires there only be a single RoboTaxi though… It’s a faulty comparison.
Tesla’s Robotaxis suck, but you don’t have to go about making questionable statistics to show that.
The statistics make sense. The metric is total num crashes divided by total miles by fleet. It does not matter if one Tesla drove 10 km, or two teslas drove 5 km each.
It absolutely does, as they could have an accident with each other. If your statistics were taken, we could simplify down to a single vehicle globally, which can’t have an accident with another vehicle as we’re just removed them from consideration.
No it doesn’t. This effect only comes into effect if there are sufficient teslas on the road that p² is comparable to p, where p is the ratio of teslas to all cars. if 5% of all cars are Teslas, then p is 0.05 and p² is 0.0025, which is a negligible probability. Teslas are much more likely to encounter non-Teslas than Teslas.