.svg){kind=icon}
Navigating the Route to Personal Autonomous Vehicle Insurance
What’s Autonomy?
Autonomous vehicles are no longer a science fiction concept like in the movie I, Robot, where Will Smith calmly sleeps in the driver’s seat without a steering wheel. Today, you can get a near I, Robot-like experience thanks to companies like Waymo, which offers fully autonomous ridesharing services with electric Jaguar I-PACEs in San Francisco and Phoenix. The steering wheel is still there, but not for long.
What Waymo offers is not something you can go and buy, though. Sure, you can get a ride, but you can’t purchase one of the autonomous Jaguars with Waymo’s software and sensors. It's not because it would be astronomically expensive (each is probably around $300K), but because selling autonomous vehicles to individuals is not something Waymo does (or can ever do). On the contrary, you can easily buy a Tesla today and subscribe to FSD software, which offers a nearly hands-off experience but isn't as fully autonomous or reliable as Waymo.
There are two worlds in the autonomy universe: commercial and personal. Both have different paths to success. The commercial model is capital expenditure and operations-heavy, costing Waymo over $10 billion to date to achieve minor deployments ($10.5 billion just from public sources; the amount spent before that is unknown but likely significant). However, the good thing is that it works! The personal model is much easier in terms of capital expenditure because people buy vehicles from the manufacturer. But the performance is harder to achieve because you don’t have the luxury of unlimited sensors and endless computing power. You must keep the costs low for consumers (e.g., Tesla’s move to vision-only self-driving).
Both models will likely work and reach scale and profitability in the long term. However, there is a curveball that only a few people talk about: liability. And this curveball is tough to hit.
Commercial vs. Personal Liability in Autonomy
In commercial autonomy, liability stems from the fleet of autonomous vehicles that can cause accidents on the road, leading to third-party bodily injuries or property damage. For example, if an autonomous vehicle hits a pedestrian and causes an injury, the company behind that fleet and its insurance would be responsible for compensating the pedestrian. There are insurance solutions on the market today that can protect autonomous vehicle fleets against all kinds of accidents. For instance, Koop provides insurance products to autonomous vehicles and robotics companies with a seamless tool in 20+ use cases globally:
- Commercial AV Liability for companies that provide autonomous ridesharing or trucking services. This insurance acts as a first line of defense for on-road accidents involving autonomous vehicles.
- General Liability for product-related mechanical or engineering malfunctions that can cause fleet-wide issues with performance, suitable for developers.
- Autonomy Errors & Omissions and Cyber Liability for software-related or security protection issues that can lead to third-party damage, suitable for service providers.
The challenge arises from personal autonomy. In fact, there are no insurance products available on the market that will insure you against a personally-owned autonomous vehicle making a mistake and causing an accident. Why? Well, because there are no personally-owned autonomous vehicles just yet. That might change soon. Tesla is getting ready to introduce the first personally-owned autonomous vehicle on October 10 this year.
It will be a milestone event in the industry, and it poses a reasonable question: how are we going to insure self-driving vehicles that consumers will own? Long story short: it won’t be easy for consumers or AV manufacturers. However, those who solve the insurance challenge will reap huge benefits.
Insuring your Personal AV
Typically, liability insurance follows responsibility. It has been that way since the existence of modern transportation. Humans have been responsible for operating vehicles, from our own cars and trucks to heavy equipment and trains. Now, that has started to change as autonomy becomes the driver.
Let’s hypothesize and use Tesla as an example. You order a Tesla robotaxi, dubbed the Cybercab. You purchase the vehicle, which means that Tesla has nothing to do with keeping it on its balance sheet. The vehicle is rumored to come with no steering wheel, but it will have to be available for people to drive manually (so likely foldable). Whether it will come with FSD pre-installed and included in the purchase price is unknown, but since the Cybercab will be a robotaxi, there might be some FSD incentives or discounts.
Now, when you drive your Cybercab, there are two modes: manual and driverless. In manual mode, you are the driver, so you are responsible for the liability. If you hit someone, it is on you and your insurance company (e.g., GEICO or Progressive). This is the easy part.
When you switch to driverless mode, you are no longer responsible for driving. The vehicle is! The challenge here is that so far, all Teslas have come with a steering wheel and an explicit disclaimer that a driver is responsible for everything, even if FSD is turned on. So if you get into an accident in your Model 3 or Y with FSD on, you, the driver, are still responsible, and GEICO or Progressive will be on the hook too. It’s not impossible to imagine that GEICO or Progressive could go after Tesla after investigating an accident, but most likely, they wouldn’t unless it was a very costly accident.
Now imagine there is a foldable steering wheel (note: both the Cybercab and FSD will need to get some form of regulatory approval), and you are running FSD without the steering wheel. You no longer have control of the vehicle; it’s all Tesla’s software. So if you get into an accident in a Cybercab on FSD without the steering wheel, your GEICO or Progressive won’t cover you anymore because you are no longer responsible for driving. What do we do here?
Technically, Tesla is fully responsible for the performance of the Cybercab in autonomous mode. Tesla can use its balance sheet to pay for the losses related to Cybercab accidents (not advised) or use its corporate insurance for that (but corporate insurance can be used up to a certain limit known as the “aggregate limit” for product liability and errors & omissions, whereas traditional auto insurance has no aggregate limit). This is a tricky situation because both options are sub-par: the former exposes your cash to unknown risks, while the latter might not provide enough protection. Additionally, the liability switch between manual and autonomous modes will be a nightmare to track because it will require hyper-detailed data reports from the OEM.
How are we going to insure the Cybercab and its owners, then? There is an answer. Tesla started working on something back in 2019 and likely envisioned that its prime time would come around the Cybercab launch. It’s called Tesla Insurance. An OEM taking control of the insurance in the age of automation is the most likely path to solving personally-owned autonomous vehicle liability.
OEMs Turning Insurers
Tesla Insurance is a Tesla-owned and fully integrated insurance operation (usually known as the MGA) that provides Auto Liability insurance to Tesla customers at the point of sale. For example, you go online to buy a Tesla vehicle and get offered a Tesla Insurance quote right away. Today, Tesla Insurance provides market-level coverage and leverages telematics for rating purposes, including driver behavior and the use of Autopilot features.
Tesla entered the insurance business to reduce the cost of owning electric vehicles. While this was the main goal and it gained traction among Tesla customers (Tesla Insurance now generates $500 million in premiums annually), the secret objective of the insurance venture was to prepare Tesla for the Cybercab moment.
Instead of owning the risk or using its corporate insurance, Tesla will likely introduce a new insurance product specifically designed for personally-owned autonomous vehicles like the Cybercab. Let’s dub it a “Personal Autonomy Cover,” or PAC policy.
For customers, a PAC policy will not distinguish between manual and autonomous modes. It will likely be structured to cover liability from Cybercab accidents regardless of the mode used. Why? Because Tesla will already have access to all the data coming from Cybercabs. The cost of the PAC policy, which will be determined by both human driver performance, FSD performance, and the distribution between the two, will be embedded in the cost of the Cybercab. Customers won’t even have to think about buying insurance.
On the backend, Tesla will treat manual and autonomous accidents differently. When a Cybercab is in autonomous mode, it technically becomes part of a Tesla network or fleet. In theory, if it’s a fleet, there should be a commercial insurance policy for auto liability for Cybercabs on the network.
To take it even further, when Tesla starts offering autonomous ridesharing, people will be able to lend their Cybercabs to the Tesla network and make money from it. In addition to having non-paid Cybercabs on the network, Tesla will also have paid Cybercabs providing rides. However, Tesla won’t own or directly operate those vehicles; it will be the FSD software and the network provided by Tesla.
We envision that a PAC policy will cover the full continuum for Tesla owners, from manual driving to autonomous mode without the steering wheel, to paid rideshare rides. If an accident happens, owners will use this insurance like any other insurance policy, while Tesla will manage the numbers appropriately on the backend. The limits of the PAC will be market-level, with similar coverage provided, and the PAC will endorse or include FSD and ridesharing altogether.
A huge benefit of running a PAC program will come from claims handling. What body shop will be able to repair a Cybercab? Likely no one in the near future. Tesla itself will have the expertise and supply chain to perform the repairs, giving it a significant advantage in keeping the cost of insurance down. This will be a key factor in PAC pricing, in addition to the reduced frequency of accidents, for which we have promising yet still early data.
Tesla can use part of its balance sheet to cover the risk while reinsuring most of the insurance program. Reinsurance will provide certainty for capital allocation towards the insurance venture (e.g., Munich Re is the biggest player in automotive reinsurance). In addition to offering and running PAC at the Tesla Insurance level, Tesla itself will have a corporate E&O and Cyber Liability program to protect itself in case of a widespread or catastrophe-like event.
The biggest advantage of the PAC is that it will be fully integrated into the OEM, allowing Tesla to have full control over the experience and capital. This is crucial when introducing a new paradigm of personal mobility. This has been the secret reason why Tesla Insurance was started, and it will serve its purpose well.
Main Players in Personal Autonomy
While Tesla is clearly the trailblazer, there are a handful of companies trailing right behind it. In the U.S., Hyundai has the second largest market for electric vehicles (when combining Hyundai, Kia, and Genesis). There are predictions that Hyundai might take the route of developing personal autonomy, either by itself or through its venture Motional. Currently, Hyundai works with Mobileye. If or when the transition to internal development happens, Hyundai will face similar issues with liability for personal autonomy. Luckily, there will be answers by then!
For the latest updates on the main players in personal autonomy and autonomy in general, follow the AUTONOMY LEADERBOARD by The Road to Autonomy.
