Why March 9 Could Be a Huge Day for Tesla Investors

Tesla (TSLA 1.49%) is expected to deliver critical crash data on potential full-self driving (FSD) traffic violations to the National Highway Traffic Safety Administration (NHTSA) on or before March 9.

The data is certainly meaningful, as it’s part of an NHTSA investigation. Moreover, given that Tesla has reported 14 incidents involving its robotaxis (using unsupervised FSD) since their inception in June 2025, the company and its robotaxi rollout appear under significant pressure. But just how bad is it for Tesla?

Taking an even view

The NHTSA hasn’t published anything on the extended March 9 deadline yet, so investors are relying on others’ reports on the matter. It’s important not to downplay the data because nobody, and above all, Tesla, wants traffic violations from FSD. It’s also important not to overplay the data.

First, there were reportedly more than 8,300 records left that Tesla needed to review. That is not crashes. In fact, the NHTSA identified only 58 incidents when it opened the investigation. It’s clear Tesla is struggling to deliver the data quickly. But I don’t think this means Tesla is delaying delivery.

Second, it’s important to note that this data is different from the robotaxi incident data, which is reported and handled separately. More on that in a moment.

Third, the data spans the early days of FSD to more recent days and is likely to mainly cover older versions of FSD. That might prove to be a blessing as well as a curse. For example, it won’t be good news if the updated versions of FSD still have the same issues, or if fixing one issue creates another. However, it will be positive if the data demonstrates a linear improvement.

Image source: Tesla.

Robotaxi incident updates

Context is also required when analyzing the robotaxi crash reports. They are reported under the NHTSA’s Standing General Order on Crash Reporting, which also includes data from Waymo and others.

Tesla bears will argue that the robotaxi collision rate of about one in every 57,000 miles is much worse than Tesla’s own estimate that the U.S. average driver has a major collision every 660,000 miles and a minor collision every 222,000 miles.

However, this is an apples-to-oranges comparison.

First, robotaxis aren’t clocking up miles on interstate highways; they are driving in a relatively restricted urban area. So the 660,000 and 222,000 miles figures above are not much use for comparison, because they include highway miles.

Second, it can pay to look at trips rather than miles. Alphabet‘s Waymo assumes an average trip length of just 4.3 miles. That number is significant because, by Tesla CEO Elon Musk’s admission, on an earnings call, “all it takes is like 1 in 10,000 trips to go wrong and you’ve got an issue.” Using the 4.3-mile-per-trip estimate and an estimate of 800,000 robotaxi miles, some simple maths gets you to one collision in 13,289 trips if you consider that Tesla has reported 14 accidents with its Austin robotaxi service.

Third, I covered this issue previously, when Tesla had seven collisions over about 250,000 miles, so it appears it took 550,000 miles to reach the subsequent seven collisions, implying that robotaxi is improving safety.

Fourth, Tesla publishes safety data on its website, and it clearly shows that a major crash involving supervised FSD occurs every 5.3 million miles, compared to the average U.S. driver’s 660,000 miles. While this doesn’t validate robotaxis in itself, it does demonstrate the safety benefit of FSD.

Tesla robotaxi incident data

Data on the 14 robotaxi collisions are publicly available from the NHTSA, and it is summarized below. Understanding that Tesla is obliged to report minor collisions that regular drivers are unlikely to report (for example, hitting a tree at less than 1 mile per hour), Tesla’s robotaxi collision data is actually quite impressive. 

Data source: National Highway Traffic Safety Administration.

In addition, many collisions occurred at very low or zero speed, or the robotaxi reportedly came to a stop before the collision. As a reminder, the reports do not attribute fault.

What it means for Tesla investors

The robotaxi data isn’t perfect, and it’s hard to know what might come out by March 9, the next deadline. Still, the current data is a lot better than some headlines might have you believe.