Perception on Tesla (III) - The Value of Tesla Insurance

Tesla Announces Sale of Its Own Insurance

Recently, Tesla started selling insurance in the United States, but it is currently only available for owners in California, which is one of the areas with the highest number of Teslas in the country.

Tesla’s own insurance is 20% to 30% cheaper than insurance from other companies, which I believe is one of Tesla’s most valuable businesses in the future. Let’s take a closer look at the value points of Tesla’s own insurance.

(From Tesla North America Official Website)

The fact that Tesla’s insurance is 20% to 30% cheaper than outside insurance may indicate a relationship with Autopilot (abbreviated as AP), which was discussed in a previous article. The logic is that Tesla’s AP reduces the accident rate, which ultimately affects the user’s rates.

To understand the value points of this business, we need to understand how much safer Teslas are compared to traditional cars. But does this mean that Tesla’s accident rate is 20% to 30% lower than traditional vehicles? I don’t think so. To be more rigorous, let’s take some time to analyze public data to see if this is really the case.

Improving Safe Driving Mileage

After searching for Tesla’s publicly released vehicle accident statistics on the internet, I found that they are calculated in three main modes: AP (with active safety) enabled, only active safety enabled, and neither AP nor active safety enabled. It is not difficult to see that the accident rate is lower when AP is enabled, much lower than the accident rate when only active safety is enabled, and much lower than the accident rate when neither AP nor active safety is enabled. Thus, Tesla’s vehicle accident rate is indeed lower.

(From Tesla China Official Website)

We need to carefully consider the scenarios in which these three modes are used to determine the meaning behind each set of data and which set of data has the most significant reference value compared to traditional data.

First, let’s look at the data when AP is enabled. In the previous article, I mentioned that AP is mainly used in highway scenarios, with a usage rate of up to 80%. For urban scenarios, the usage rate drops to around 50%. Therefore, the data in this mode is obtained under simplified conditions, and I do not think it necessarily has practical reference value. This set of data should be compared with data on high-speed accident rates in traditional cars to be truly meaningful.Then again, let’s look at the data that only includes active safety systems. This data should primarily come from urban scenarios, which are relatively complex environments. Most ordinary vehicles drive in urban environments, so this set of data should have the highest overlap with the scenarios that traditional vehicles drive in. Therefore, this group of data can be used to compare with data from traditional vehicles.

On the other hand, the data without active safety and without autopilot is equivalent to data from traditional vehicles.

In addition, I mentioned in my previous comments that the usage rate of autopilot in urban scenarios can reach 50%-70%. I am re-mentioning this point to correct the fact that statistics including the use of autopilot in urban areas should be included in the data for active safety only. This means that the safety driving interval mileage in the data for active safety only should be underestimated.

So, temporarily using Tesla’s data for active safety only as a comparison, a report of one accident every 2.42 million miles driven (active safety only) vs. one accident every 1.79 million miles driven (without autopilot and active safety) is reported. (242-179)/179 equals 35%, which means that the safety mileage has increased by at least 35%. This seems to corroborate with the previous statement that Tesla’s own insurance is 20-30% cheaper than others.

Lowering the rate of accidents triggered by active safety

However, the demonstration of the value proposition for this business is not over. Let us consider the compensation mode for insurance, which is compensated by the party that actively triggers the accident, such as when a rear-end collision or a cross-over of solid lines/forced lane merge is initiated.

In other words, only the accident data was presented earlier, and there is no distinction between accidents initiated actively by Tesla, or accidents initiated by other drivers. In other words, from the perspective of Tesla’s own insurance, only a portion of the compensation claims need to be compensated by Tesla.

So, let’s explore the issue of accidents triggered by active safety systems. Can Tesla’s system reduce the rate of accidents triggered actively by the driver? This will affect the commercial value of Tesla’s insurance business. Based on my experience that involves using Model 3, I have observed that Tesla’s AP and active safety strategies are geared towards conservative driving, and most of the incidents that require my intervention are due to the sudden lane-changes of other cars, which drive in lanes in a way that is different to mine. In other words, if there is an accident, it is mostly the other drivers that are liable.

This means that the probability of actively triggered accidents with Tesla should be much lower than other vehicles.

Therefore, it can be concluded that Tesla’s safety driving mileage should be at least 35% higher than that of traditional vehicles in a certain scale, and the probability of accidents triggered by active systems should be much lower than traditional vehicles. In this sense, even though Tesla’s insurance is 20%-30% cheaper than other insurance companies, its AP and active safety system still give it a lot of value.

What’s most impressive about Tesla’s business is the ability to reduce costs by technological means, while also earning profit.

This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email