Who Insures the Personally Owned Robotaxi Fleet?

Executive Summary

In this episode of The Road to Autonomy, Grayson Brulte sat down with Sergey Litvinenko, Founder & CEO of Koop to dissect the upcoming commercialization of the Tesla Cybercab. They discuss the rumors of a February network opening and the anticipated production ramp-up in April 2026. The conversation covers the evolution of insurance, debunking the myth that LIDAR is required for underwriting, and how the industry is shifting toward data-driven risk assessment.

Furthermore, Sergey explains the emerging business model of “personally owned fleets,” where individuals form LLCs to operate multiple Cybercabs, and how liability will be split between fleet managers and Tesla’s software.

Key The Road to Autonomy Episode Questions Answered

Key The Road to Autonomy Topics & Timestamps

[00:48] Tesla’s 2026 Timeline: April Production and the February Network Rumor

Sergey highlights that 2026 will be a pivotal year for Tesla, pointing to public plans to ramp up Cybercab production around April 2026. He also discusses rumors suggesting that Tesla may open its network as early as February, allowing users to register vehicles for self-driving operations. This process will likely involve registering vehicles as commercial entities rather than personal vehicles to comply with state regulations.

[04:44] Debunking the Myth: Why LIDAR is Not Required for Insurance

Addressing common misconceptions, Sergey explicitly states that LIDAR is not a requirement for securing insurance. He explains that modern underwriting focuses on real-world performance data rather than specific hardware stacks; if the data demonstrates safety, the risk is insurable regardless of whether LIDAR is present. He notes that the insurance industry views data as “king,” making Tesla’s camera-based approach viable if the metrics support it.

[08:21] The Rise of the Fleet: Moving from Personal Ownership to LLCs

The discussion shifts to the economic structure of the Cybercab, with Sergey noting that it is impossible for Tesla to carry tens of millions of vehicles on its own balance sheet. He predicts a shift where individuals form LLCs to buy multiple vehicles, effectively becoming small commercial fleet operators. This allows Tesla to scale and collect software revenue without bearing the heavy capital expenditure of owning the physical assets.

[13:20] Geographic Constraints and Geofencing for Fleet Owners

As fleet owners deploy vehicles, they will likely set specific geographic boundaries or geofences for their operations. Sergey explains that these decisions will be numbers-driven, based on logistical factors such as the distance to charging stations, the ability to retrieve a vehicle if it breaks down, and the location of maintenance depots.

[23:09] The State of Underwriting: Capacity and Understanding in the Market

Sergey admits that there is currently not enough capacity in the insurance market for autonomous vehicles because the traditional industry does not fully understand the sector yet. While reinsurers are showing appetite because they are mathematics-driven, primary insurers are lagging behind. He anticipates that the 2026 launch will hit the industry “like a snowball,” forcing carriers to adapt quickly.

[25:28] Koop’s New Product: Primary Liability for Autonomous Vehicle Fleets

Koop is developing a specific insurance product designed to be the “first line of defense” for autonomous fleet owners. This policy covers the vehicle across all states—whether in manual mode, autonomous mode, or parked in a garage. If an accident occurs, the policy pays out, and if the investigation reveals a software error, Koop will then subrogate the claim against the developer (e.g., Tesla).

[33:13] Liability Split: Distinguishing Between Software Errors and Maintenance Failures

Liability in the future will be shared contractually between Tesla and the fleet operator. Tesla will retain liability for accidents caused by software mistakes, treating it as a product liability issue. However, fleet owners will carry liability for maintenance-related risks, such as accidents caused by bald tires, uncleaned sensors, or failure to update the vehicle’s software.

Full Episode Transcript

Grayson Brulte: Sergey, sir, it’s great to have you back here on the road to Autonomy and thanks for being the world’s greatest sponsor, which you are, Koop you sponsor this week in the Autonomy Economy. So thank you for allowing us to publish our opinions and analysis. And sir, you have great insights into autonomy, into fleet management and insurance. And so when I called you up and say, Hey, Sergey, I’m starting to see the Cybercabs, seeing ’em in Austin. I’m seeing them in Palo Alto. And I said, well, I know someone that’s got a really good, insightful opinion. What are your thoughts as all signs point to Tesla getting ready to ramp up and commercialize the Cybercab.

Sergey Litvinenko: Thank you Grayson. Always pleasure to be back and, uh, yes, um, all eyes on Tesla 2026 is gonna be a huge year for them. Uh, first of all, we know publicly that the plan to ramp up the production of Cybercabs around April, 2026. Um, I assume they would not have done it without having conviction about where their self-driving, um, strategy is headed. And, uh, as you noticed, um, or as you mentioned that there are Cybercab spotted on the streets, uh, probably testing before the, uh, production goes into ramp. And, um, uh, I’ve heard rumors from. From Koop, obviously that, uh, um, people started inquiring with us about insurance and from some of those conversations we learned that people, um. Think that Tesla is gonna open up the network in February. Now, we don’t know exactly what it means to open up a network, but we think it means that, , you’ll be able to put your Tesla on the network for self-driving operation as soon as February. It may be you’ll be available to limited number of people, but that’s what the rumor is.

Sergey Litvinenko: And supposedly there is a whole registration process that you have to go through. You obviously have to go through. Each state as well. , And, because technically the vehicle will be classified as a commercial vehicle, not just personal vehicle. And I think that there is at least 10 instances where people are thinking about doing that. And, um, all of them said more or less the same thing, that February seems to be the timeframe when the network supposedly will go live. Um. I think that’s huge, if that’s true. Also the timeline might be different, but let’s assume it’s February if it, if it goes live in February. Then the production starts in April. , I think by the end of the 2026 we’re gonna live in a very much different world where Tesla’s just gonna flood the market with their own fleet and also with third party owned fleets, which, uh, in my view is gonna be huge. So, uh, that’s the rumor and I think it has huge implications for the self-driving industry, especially when it comes to the partnerships and how the ecosystem will develop. And, uh, if Tesla gets into the partnerships game, that’s gonna be game changer for lots of vendors in that space.

Grayson Brulte: It’s gonna be a game changer as we look to analyze the potential deployment of the Tesla network. Should we follow the, and look to markets where Tesla currently has? Uh, an insurance license. Should we look at those markets to be the potential markets? Because you clearly stated it’s going to be a commercial application. As we’ve seen from Uber and Lyft, the insurance dramatically changes.

Sergey Litvinenko: Correct, and I think we should double click on insurance because, um. Tesla does have a personal auto insurance product in multiple states. Not all of them, but in multiple, I think in 20 or so. , But I, the question is, will the product be sufficient for the autonomous vehicle operation? I think the answer is no. And there are two ways how insurance can evolve for Cybercab. It’s either can be self-insured or you have to have a separate product for that separate policy. Now, can we use insurance as a proxy for where Cybercabs are going? Probably yes. Um, but we know for sure that it’s Texas and Bay Area. Uh, I honestly think Tesla will likely, uh, go to the biggest markets first. Similar to how maybe Waymo has scaled. If you look at the Waymo’s expansion, it’s clearly the biggest markets. So it’s California, Texas, Florida, New York, and then they tapped into some others. I believe they’re doing something Illinois, , Arizona and others. So I think Tesla will follow the similar path. So insurance will not be kind of the guiding principle for whether select to deploy, but insurance is gonna be a hundred percent. The bottleneck, how they’re gonna deploy, whether they can deploy at all, and that probably deserves a whole separate rant.

Grayson Brulte: you bring that up. ’cause on Autonomy Markets that I, I host of Walter Piecyk, we’ve get some interesting comments from time to time and, and one of the feedbacks we get, you cannot launch a self-driving car without robust sensors. You must have lidar, ’cause insurance is not gonna insure it unless you have lidar. You run an insurance company. Can you debunk that for us please?

Sergey Litvinenko: LIDAR is not a requirement for insurance that I can tell you a hundred percent. And I think the reason why some people are talking about this, there might be insurance underwriters on the market who got an idea that LIDAR works or you need lidar, uh, to be at some level of safety. That’s not how we operate. We operate differently. We look, uh, at the real world data and uh, that’s what informs our underwriting decision. So if you have lidar or don’t have lidar, if the metrics show good performance, so that risk becomes under rideable for us. So, uh, I think this whole notion of lidar, I. As far as I know has been disproved. And you can see there are some players on the market, which went from having a very high valuation to now closing down and some OEMs pulling out of the LIDAR deals. Um, I think that’s telling and um, with the imminent launch of Tesla, which as we know Tesla doesn’t use light, lidar. I think we’re gonna see a situation where data becomes the king, and if data is good, then it’s good to go. That’s what insurance, by the way, has all been about in the first place. It’s been about data. So if I can have some new technology and I know which data I’m looking for and that data is good, then that risk becomes insurable. So it doesn’t necessarily mean what kind of specific things you have in your tech stack. Um, that’s secondary in our view.

Grayson Brulte: What type of data are you gonna look for? When an individual buys a personally owned Tesla, they buy a personally owned Cybercab, what data are you gonna look to when you look to potentially underwrite those vehicles?

Sergey Litvinenko: we’ll talk about the structure. Uh, we need to talk about whether it’s gonna be the individual or fleets or Tesla itself. In terms of the data, um, we are looking at the behavior of the vehicle. , If you look at the, probably last 30, let’s say last 20 years of automotive insurance, , of auto insurance, both commercial and personal, one thing that we were able to do we’re able to connect driver behavior to the claim outcomes. So, uh, it’s been proven that the worse you drive, the higher chance you’re gonna get into an accident, um, that’s been proven okay? Or the worse you drive, the higher the chance that you’re gonna become a threat for other people who drive well on and you share the same road with them. So, uh, we take the same notion to self-driving cars. So if we can, uh, drive the difference between how humans drive. Versus how AVS Drive, that informs a lot how we make underwriting decisions. Um, Waymo posted their safety stats out there. The difference is clear. It’s 80 to 90% better performance. Um. And humans, Tesla has been posting statistics about their autopilot and FSD performance. It’s the same story. So we carry that over into our underwriting program, which passes the benefits, um, these benefits to, um, self-driving, car developers and operators eventually.

Sergey Litvinenko: Um, the metrics that we look at, um, are all about the behavior. So you can think of the worst thing that happens is an accident, but then there are. Behaviors and types of behaviors leading to that, uh, to that accident. It could be harsh braking, harsh acceleration, harsh turning. It could be near misses, which by the way are very telling of the potential accidents that you might have. And then you do it across different, uh, parts of the city or parts of the, uh, operating domain. And then there’s a lot of data to play with, but it’s like a pyramid, you know, there are harsh accidents. Which are the real crashes. And then there are, are near misses, and then there are like smaller behavioral accidents and they all correlate to each other. So that’s our approach and I believe that’s the approach that other people take. Um, but that’s the data that we’re gonna mine for across all the use cases, be that Cybercab or some other OEMs or operators.

Grayson Brulte: And since it’s, it’s my belief that a lot of these Cybercabs are going to be deployed in fleets. Where you’re gonna have an individual, perhaps pull some money together with friends, they set up an LLC or some corporate structure, five vehicles, 10 vehicles, and you have the larger fleet owners. Perhaps we see it in trucking. You might have an individual who owns two or three of them. How is that gonna change underwriting since you’re putting an entity in the middle of it?

Sergey Litvinenko: I think, um. So first of all, let’s lay out the, um, three different scenarios, um, for Cybercab ownership, um, and then we’ll talk about fleets. So the first scenario, which is what’s happening right now, is that Tesla owns and operates the fleet. So Tesla has the uh, uh, uh, robot network. They, so it’s like fleet management, uh, tool that they use internally to match supply with demand. They have a customer facing app, then they have the Cybercabs and then obviously have the software. And they do everything in a vertically integrated way, meaning. They own the vehicle, they carry it on the balance sheet, they provide the service, they collect the money, and then they also, uh, maintain the vehicle. So they clean it, charge it, update the software, et cetera. Um, that’s how it’s gonna be. And it’s always gonna be there because Tesla will probably want to test out their latest software. Um, they might want to test out different markets, so they’re always gonna be doing that. The question is, is that the model that they’re gonna scale with?

Sergey Litvinenko: So, um, just think about the sheer number of vehicles out there. I think in the US it’s hundreds of millions, and then it’s like 90 million new vehicles sold every single year around the world. It’s impossible for one company to own all of those, all those vehicles, right? Um, so I, I think the largest fleet in the world might have just a little bit. Maybe close to 1 million vehicles and 1 million self-driving cars are not gonna cut it for the, just for the us. Forget about the rest of the world. So we’re talking about tens of millions of vehicles. Okay. Now, will anybody want to carry that on their balance sheet? I think the chance of that is zero. Not Tesla, not Waymo, not anybody else. So what’s the second best model for Cybercabs? Well, the second best model is to do what Tesla has been doing really well in the past 15 years to sell those vehicles to individuals, to me, you and everybody else. Now, um, I know you own Tesla, you know, and you drive it every single day for personal use, but, um. Will you as an individual be able to make money from the Tesla by putting, putting it on a Tesla network? Um, it looks like they circulated this idea, but technically I don’t think it’s gonna work like this.

Sergey Litvinenko: Most likely what’s gonna happen is that you, Grayson, you’ll probably say, Hey, if I can put one vehicle on a Tesla network and make money from it, why can’t I put two or three or five or 10? And now if you’re thinking about one or more vehicles, you’re already becoming a fleet. Now you need to form an LLC and you’re gonna be treated as a commercial fleet no longer as a Grayson’s Tesla. And then that LLC will have to transact with Tesla to put those vehicles on the network. And I think that’s how it’s gonna evolve. We’ll likely see. A lots of smaller fleets popping up all over the country. Some people owning like literally one vehicle, some people doing maybe 1,000, eventually a thousand. And that way Tesla can scale without carrying all the CapEx, but collecting all the software money. And that’s gonna be a game changer, not only for Tesla, but also for. People who want to make money from autonomy. And that’s where I think it’s gonna open up huge opportunities for third party partnerships for Tesla. And it’s probably gonna be the first instance where we’re gonna see Tesla share its, uh, economic value, uh, with other people rather than vertically integrating. And that’s gonna be huge. So, um. That’s where everything is headed in my view, and I do not think is gonna be personal ownership in the future. There is gonna, there are gonna be personally owned fleets, like LLC owned fleets, but I do not think that individuals will have to own Teslas and, uh, put them on the network.

Grayson Brulte: The fleet aspect is interesting. So let’s just say I’m a fleet owner. You Sergey, you’re a fleet owner. Let’s just do simple math. I own a hundred vehicles. You own a hundred vehicles. You haven’t deployed in Northeast. I haven’t deployed in the southeast. Do you see a point where the fleet owner will set a geographic OED? Okay. I don’t want you going out of this state. I don’t want to go more than 300 miles away from base. Do you see stuff like that potentially getting implemented into this fleet model?

Sergey Litvinenko: Absolute. Um, I, I don’t know if you’ll be able to set a custom geo fence or it might be, uh, zip code constrained, but that’s how it’s gonna be. You mentioned, uh, trucking earlier, so I know that there are many, uh, independent fleet owners, operators where they operate maybe on one or two routes. Okay. And they happily operate with big trucking companies, um, along alongside the same routes. So probably will be a similar thing with, so with, with Cybercabs that you choose a location and then you, um, set a certain perimeter that you operate with them. Um, and you know that decision can be driven by multiple factors. Uh, first of all, okay, how far do you plan to go knowing that you have to charge the vehicle? Okay. Or let’s say if something happens to the vehicle and you have to retrieve that vehicle, uh, how far are you willing to drive? Okay. Or let’s say it’ll also mean, um, that if you, let’s say, have a, uh, depot in Jersey City, maybe I will not be able to go to, let’s say, upstate New York because it’s gonna be too far. Okay. Uh, and then I don’t want, let’s say if my passengers want to book me from Jersey City to upstate New York, I’ll have to set a limit on how the ride can be. And, uh, I’ll just not be able to do it.

Sergey Litvinenko: Um, or I’ll have to figure out where I can charge the vehicle there so it can come back. So it’s gonna become very, um, very, uh. Numbers driven, you know, and people will decide, okay, here are the parameters I don’t wanna violate for my operation, and here’s how much, um, area I can cover and how much money I can make from it. So it’s gonna be fascinating. The question is, is it gonna be done manually or is it gonna be a product like that where you can automatically figure out which area you can serve? And maybe it can be done by Tesla or by some third party software providers? It’s, it’s, it’s, we, we yet to see. You’re a hundred percent right that they’re gonna be constraints and people will be slicing and dicing the United States in terms of who, who got what part of the self driving car operation?

Grayson Brulte: So then do you envision the vehicles only being available on the quote unquote Tesla network? Or do you see these vehicles potentially going on third party networks?

Sergey Litvinenko: great question. I think Tesla will run on Tesla, and. Other, other OEMs probably will be able to run on Uber and Lyft. So let’s say if I am a fleet owner operator, really. Um, let’s say I have the financing. So I have a few billion dollars and I want to run an autonomous vehicle business, and I want to make money by giving passengers rides. Um, I go and buy Teslas, okay? Then Tesla runs on a Tesla network, but I also want to go and buy Waymo’s. And Waymo will run on Waymo network or maybe on Uber. Maybe on Lyft. And then let’s say I want to go and do somebody else, like wave or emotional. And, um, it would be great to have one place where you can just run everything. Maybe you run everything in Uber. You know, um, I think it’ll be great, but I just don’t know how Tesla plans to work with Uber. Um, I wouldn’t be surprised if Tesla becomes, um, a, uh, closed ecosystem and that you can either do Tesla or something else, or it’ll require to have different apps. Um. We get to see. So maybe there is an opportunity to develop a, um, a tool that will actually aggregate all of those in one place where you can do fleet orchestration across multiple autonomy platforms in one place, and then you can automatically connect to Tesla Network, Waymo, Lyft, Uber, et cetera. Um, and you can actually kind of, uh, switch vehicles from one network to another. That would be pretty cool. I don’t think a product like they’ll, like this exists. It’s not needed right now, but I give it a couple years. They, they can be an opportunity.

Grayson Brulte: I see the Tesla vehicles operating in a closed network, but. I’m not a lawyer, so I can’t give legalities on it, but I mean, I wonder if there are certain legalities. If you legally buy a product, you can only operate on a network, not an antitrust, an attorney. So I can’t give an opinion there, but I think that’s ultimately where Tesla would want to go. And I ultimately believe that individuals that buy the Tesla vehicle, the Cybercab in this case, will want to deploy on the network and then. What you mentioned earlier about the limitation where the vehicle can go, I’m sure that’s gonna be a setting inside of the Tesla app. When you put your, your vehicles on there, and you’ll probably introduce some contractual agreement with Tesla, let’s called Tesla Rideshare, whatever they’re gonna call the entity and it’ll tell you what you can’t do. Let’s look at certain areas of the United States. Let’s look at Westchester County to Fairfield County. So you look at the, uh, Bedford Greenwich line overlaps all the time, o over there from one state to another. How do you manage that from an insurance perspective? Do you have to ensure that that vehicle, if you’re operating in that area, is able to legally operate in Connecticut and in New York?

Sergey Litvinenko: Great question. I assume that there’s gonna be either a federal framework or a state by step framework that allows self-driving cars to operate in the first place. Let’s get out of the way. I think it’ll be done. Okay. Now, insurance wise, right now insurance is, uh, sliced up. Across all 50 states, there are different requirements for commercial fleets, so you might be operating in one state as soon as you cross to the other state. You need to make sure that you satisfy both requirements. Usually requirements come down to having an insurance policy from up to a certain limit from a carrier that has certain rating. That’s about it. Okay, so the state just wants to make sure if an accident happens that everybody’s protected. Um, and if you have multiple states and have multiple requirements, you satisfy the highest requirement, you’re good to go. Now the question is how the heck you’re gonna satisfy the requirement if it’s self-driving car? Okay, so, uh, the car is literally, uh, it might have a wheel, might have no wheel. And then, uh, for you to ensure it, it means that there has to be an insurance company that’s willing to take the risk. Now, there are two ways this can go. The first way is that Tesla says, look, we are going to provide you with an insurance product, okay? Or we’re gonna self-insure and test self-insure and extend that self-insurance to our precious, uh, fleet owners.

Sergey Litvinenko: Now it’s possible, but that means that Tesla will carry all that risk, risk on a balance sheet. Um, will they be able to carry, uh, a risk for millions of Cybercabs simultaneously? Um, great question. I think the answer is no. I think it’s gonna be way too much risk, uh, because don’t forget, auto liability is not, is not capped. You know, if you accidentally incur a million accidents, um, at the same time, you know, it will just stack up without the ceiling. It’s not like product liability where there is usually cap or the aggregate limit. So I think most likely. What’s gonna happen is that there will be new products on the market and that those products will be specifically built for, um, autonomous vehicle fleet owners and operators. And, uh, that product will take, uh, uh, into account the fact that the driving is done by software. That the software performs extremely well. And most of the risk is actually gonna be in some miscellaneous things like, hey, you know, unpredictable weather, or, uh, there was an error in maintenance, or some really, you know. Acts of God that you cannot think about. Uh, and that auto, auto liability product, um, uh, for autonomous vehicle fleets will be, uh, very well priced, meaning that most of the risk from human driver is gonna be, uh, taken care, uh, by the software. So it’s not gonna be in your policy. And let’s say if something does happen due to the software error.

Sergey Litvinenko: It’ll be subrogated against Tesla or other OEMs and whoever provides that insurance policy will have to figure it out with Tesla. And now you have a situation where the autonomous vehicle risk can be spread out by many different insurance providers in different parts of the country. And that becomes much more manageable just how you let other people buy Cybercabs and, you know, make. Benefit from those Cybercabs, you will let other insurance companies ensure autonomous vehicle fleets, um, across, uh, a huge capital base, which will not be concentrated within one OEM and can take that OEM down if something, if something happens. So, um, from the insurance perspective, that’s probably how it’s gonna happen. And I have, uh, a ton of confidence that, um, each state will have no problem with such an insurance product. And then such an insurance product will satisfy all the state by state requirements. So, , to get back to your point, if you have a Cybercab operation and you go from one place or from one state to another state or to multiple states, , there is a solution where it’s not gonna be a problem. And, um. I do not think that, I mean, I really hope the regulations will, will catch up and insurance will catch up. Um, but when it comes to having an insurance product to meet all those regulations, we will have that.

Grayson Brulte: How much capacity, or I’ll ask you. This way. What’s the appetite in the underwriting world today for autonomy? ’cause you see it just on a personally owned Tesla, which I own with FSD, those rates go from here to here to here, and there is no consistency. And, and so I’m the weirdo that calls and asks questions, why? Oh, to replace the, the technology. I was like, okay, you just fundamentally don’t understand this. Is there capacity, is there appetite? Just based on what we’re seeing in the, in the personal car ownership underwriting market.

Sergey Litvinenko: I would say there is not enough capacity and. The answer is I don’t think the insurance industry fully, um, understands the, the self-driving car space. So, I mean, originally the insurance industry got into that space by insuring projects, you know, one time project where you have, you know, a certain location, it’s super geofence. It’s primarily for r and d purposes. And then when we got, . Uh, Waymo and, and Cruise, which no longer with us. , We entered the first, um, commercial deployment and that commercial deployment was, uh, self-insured. Okay. So, and you know, it works well. I mean, if you look at Waymo, the results are pretty good. Um, now we’re about to enter the third phase where autonomy actually goes into the wild and there’s gonna be different fleets owning and upgrading. So driving cars, so. For that specifically. I think the insurance companies are not prepared yet. And the good news is that I think it’s pretty clear how the product should look like and what it should do. , I think the insurance industry in general just lags, um, their thinking and their momentum. , As I mentioned earlier when we started the podcast, 2026 is gonna be huge year for Tesla and it’s gonna hit a lot of people like a snowball that, oh my god. This happened, you know, we need to do something about it. And, um, I’m really looking forward to the moment. That’s something that we’ve been working for the past couple of years, and when that happens, I, I really hope we’ll be able to help lots of those Cybercab owners and operators with, with a good insurance product that’s gonna clear lots of requirements for them. , So capacity is always an issue, but I think that eventually we’ll catch up. Um, and obviously there is a huge alpha in between now and when the rest of the market catches up, if you know what I mean.

Grayson Brulte: a big time. So you’re Mr. Insurance and you are CEO of Koop. What does the product look like, because I know a little birdie told me that you’re designing something over there, Sarah. So what’s the product look like?

Sergey Litvinenko: the product that, um, we are working on is the product that will allow autonomous vehicle fleet owners and operators to clear insurance requirements, and that is the product that will ensure. , Autonomous vehicle operation on a primary basis, meaning that if something happens, whether you are in a manual mode, in a self-driving more mode, or in a garage, that that insurance policy will be your first line of defense. Okay. And then once we know what happened and we have the data, we’ll be able to figure out, hey, is it actually the liability of the fleet owner operator, or is it the liability of the autonomous vehicle developer and developing on whose liability it is? It’ll go the the certain way. If it’s autonomous vehicle owner operator, we pay the policy. Okay. If it’s the, um, developer, then we go and subrogate subrogate the developer based on the common investigation and. Settlement that this, this, this, this is what happened, and now that goes onto their product liability, uh, insurance tower. Um, so I believe that pro and that product, by the way, actually will be, will provide a ton of convenience to autonomous vehicle fleets because it’ll be, uh, very well priced, it’ll be very convenient to get and whenever claims happen, uh, is gonna be. We’ll be able to pull all the necessary data to make a decision really, really fast.

Sergey Litvinenko: And what we’re gonna see is that we’re gonna see that the premium volume is not gonna be as much for traditional commercial autos because there is less risk. Okay? But that’s totally fine. Uh, what we’re gonna see is extreme profitability because the, um, based, already based on our experience and based on the market data that we have out there, uh, there’s gonna be way fewer losses. And that means for insurance that there’s gonna be way more profit. Then the profit can be further used to even further improve the rates for autonomous vehicle owners. Um, so that’s the product. I cannot share more, but I can tell you that this is gonna be, , a very important product to commercialize self-driving at scale in the United States regardless of what the regulation comes out to be.

Grayson Brulte: Well, when the product’s public, you’re coming back on so I, I won’t push you, but I’ll just say when the product’s public, you’re coming back on. The profitability thing is really interesting and you bring up a valid point because I believe we are getting to a point in society. Or a a, a small portion, soon to be a larger portion. And eventually a very large portion of the population will not wanna drive anymore. I don’t drive my Tesla drives me and I don’t really want to drive. And then obviously, you know, I think when you don’t drive for long periods of time, your motor skills from driving decline. So you’re, so we’re gonna get to this really interesting point. So you’re eliminating, let’s call ’em early drivers, that historically, let’s look at underwriting tables. Very bad drivers, very risky, higher premiums. And then at the end of life, older individuals, cognitive issues, bad drivers. Do we get to a point then where, where this underwriting becomes, uh, more profitable? Because you’re limiting two of the riskiest pools, and I forgot teenagers, but, so let’s put them in there. But you’re eliminating three large, risky pools.

Sergey Litvinenko: absolutely. A hundred, a hundred percent. So, um, as we discussed earlier, um, human behavior is predictive of, uh, car accidents and you just described the categories of people that tend to have risky behavior. So, um, with software just eliminates that. Okay. Um, and that’s what’s gonna drive the insurance underwriting performance. And, uh, we do think that the profit underwriting profitability is gonna be amazing, but it’s gonna come at the expense of lower premiums. Um, because you cannot, if there is lower risk, you cannot charge the same premium. Okay. It’s gonna, by the way, it’s gonna be a huge problem for insurance companies because insurance companies, um, are built in a certain way where you have a premium base. Okay? Let’s say it’s. One, or it could be a million, billion, trillion, whatever. And then a part of that premium basis used to pay losses. Okay? And historically it’s been, I think 70, around 70% across property and casualty. And then the rest of it is used to cover expenses. Okay? And they pay. They have to pay their bills, they have to pay their people, they have to buy software. Now what happens? If the losses become so great, so, so much better that you no longer have your original premium base, well, the premium base shrinks and now you have less money for yourself and you can pay as, uh, for as much people as you used to have. And the, um, model becomes challenging. That’s why I feel like, uh, it’s a lot, uh, better to start like a new insurance program or a new company that’s gonna go after the autonomous vehicle space than having those traditional insurance companies try to, uh, reinvent themselves around, uh, way lower. Premium base and way lower excellent rates. Um, I think this is just gonna be challenging for them. But on the flip side, they have a ton of capital. Okay. And I, I’m pretty sure they wouldn’t mind deploying a part of that to a new asset, like self driving cars. So, um, I think, I think what it means is I, we are going to see much better profitability. It’s gonna be lower premium, um, early on. Uh, we will not have big insurance companies make some. Aggressive moves in that space, and by the time they do, there’s gonna be a ton of alpha that somebody will capture. Yeah. I I, I, I think it’s gonna be Koop

Grayson Brulte: Well, I think you’re right. Now I’m gonna break down the stack. Where do reinsurers come into the picture are? Do they like autonomy? Do they not like autonomy? What are their thoughts and where do they come into the picture?

Sergey Litvinenko: as far as I know, reinsurance do like autonomy. If you look at top three reinsurers, that’s mini Swiss screen Hanover. They all participate in autonomous vehicle risks and automotive risks more broadly? Um, I actually think there is more appetite from reinsurance than insurers because on the primary insurance side, you, you need to build this whole infrastructure. You need to have a product, you need to have, um, data sharing capabilities. You need to figure out how you’re gonna sube against. Very technical companies such as, um, Tesla, and not a lot of people can figure that out. But on the reinsurance side of things, they more, they’re more numbers driven, mathematics driven. And when I gave, give them the same spiel that I gave to you about how, um, the frequency and the severity of losses is gonna change and how it’s gonna make sense. Um, they like that. Um, so I actually have. I’m feeling pretty confident about the reinsurance markets that are gonna appear for this type of risk, but I think I’m less confident about primary insurance doing the same stuff that we want to do. That’s why we exist.

Grayson Brulte: You do because there is going to be a need in the market that you’re gonna fill in. I’m, I’m curious if we’re gonna stay in the liability here for, for a minute. When Tesla deploys the Cybercab, which we do not have a filing line from NHTSA for a part 555 exemption yet. So for the iso, no, you cannot deploy those vehicles without the, the exemption, without a steering wheel and pedals. But at some point they will get the exemption and Cybercabs will be on the road without a pedal in steering wheel. Sorry, pedals plural. I forgot that. W where does that liability gonna live? Is that going to live at, at Tesla? And I’m asking that from a fleet owner’s perspective. Let’s just say I buy a hundred of them. I, there’s no way for anybody to manually take over. Where’s the liability gonna live?

Sergey Litvinenko: The liability is going to live. Um, the liability is gonna be shared by the fleet owner operator in Tesla, and most likely it’s gonna be contractually driven. So it’s gonna be a piece of paper that draws the line, but. I think it’s pretty safe to assume that if a Cybercab is in operational mode and then something happens, um, and it was the, the software is mistake. That is gonna be Tesla. Okay? It’s gonna be their liability because their software is, is, is, is their product. But it’s also possible that um, the vehicles that are. In possession and custody of the autonomous vehicle fleets might not have been properly maintained. Maybe they were not cleaned, maybe they, uh, didn’t check the vehicle vitals, maybe they didn’t upgrade the software. Um, it’s also possible that some of the accidents that happen will fall within fleet owner operator. Now the question becomes, how does you know liability? There’s gonna be a hundred percent the separation of liability. And that’s where it becomes an issue because, well, to satisfy state’s requirements for insurance, you need to have insurance. You know, so you need to have a product that’s gonna go after both of those liabilities as the first line of defense, AKA, the stuff that we talked about, that we working on. And then after an accident happens, you’ll figure out who to go after who. Okay. And, um, I think, I think that’s what we’re gonna land on. Um, so fleet owners, operators are not gonna be fully indemnified. They’re gonna carry liability, but it’s gonna be only a portion of the liability. The rest of it will be within Tesla, which we’ll subrogate against if they make, if something happens and, uh, it’s their fault.

Grayson Brulte: You bring up a. Very interesting, very valid point around fleet management. ’cause you’re saying that, Mike, well what happened if somebody didn’t check the tires? Seriously, you know that that can cause deflated tires hole in the tire. It, it could lead to, to a crash point blank. There’s, there’s plenty of data from DOT and Nitsa to back that up. How do you see the fleet management angle working? And is that, and who’s, who’s gonna own it? Who’s gonna operate and how are these insurance policies gonna overlap? Because. I could see a lot of quote unquote risk in that part of the stack.

Sergey Litvinenko: already today we can see Fleet Manager is appearing. , There is Move Av, there is move.io. Hertz, I think made some public statements about doing something there. I think Lyft acquired a company that is doing fleet management, uh, who will probably see some, , I don’t know, uh, new companies appearing to do fleet management specifically for self driving cars. And they’re gonna do garage. They’re gonna do charging, cleaning. They’ll have to figure out insurance a hundred percent. They’ll have to figure out compliance. Okay. Like. There are actually compliance rules for operating commercial fleets in different states. That’s a whole different thing. Um, the, and, uh, they also have to figure out how to obviously, uh, integrate with their fleets with a fleet management system, especially if they have multiple, uh, fleet, uh, vehicles and developers, um, under the same roof. Um, I do think that. Fleet, fleet management companies, some will own those vehicles, some probably will attract a third party, uh, finance partner for it. It can go both ways. Uh, I think if you own them and then you provide the fleet management alongside with it, that becomes extremely attractive, especially for AV developers who can have that. One stop partner for everything. Um, but at the same time, you can have, uh, different, um, financial companies interested in owning autonomous vehicles as, as, as real estate. And I can clearly see that. Um. Let’s say, um, you have a fleet management company where they are getting paid maybe as a percentage of revenue plus some flat fees or something like that. And there is a third party owner who’s not involved in day-to-day operations, but they just collect the check, um, from the, uh, from the revenue that the vehicles generate. Um, I think we’ll see both and extremely attractive opportunities for both. For the financing partners and for the fleet managers. So, um.

Sergey Litvinenko: And then if you think about it further, let’s say when you get to certain volume and you have millions of avs on the road and it’s predictable how they operate and there is a continuous revenue stream, then you can securitize it in a similar way how you securitize real estate. And now we off to the races and it’s the new. autonomy economy uh, exciting stuff. Uh, but yeah, I think fleet managers, fleet managers are great. Uh, there is a ton of opportunity whether they own that or do not own that. And, uh, I’m super bullish on that space.

Grayson Brulte: Thanks for plugging the autonomy economy, so thank you for for that plug. You’re right, I could see some sort of real estate investment trust, common nose reit, where the real estate’s packaged. With the vehicle or some, some sort of financial product there. Thinking down the line, this is hypothetical, but I think a lot about this. Tesla manufactures a vehicle they and build the vehicle, uh, the software in-house. They own the supercharging network. Do you think, this is hypothetical. At some point there is the potential for Tesla to launch a fleet management business to manage the Tesla network.

Sergey Litvinenko: I think they might already be doing this. So I, I’m, I’m pretty sure that, um, the imminent launch of Austin and Bay Area, um, implies that there is some fleet management, um, component to it, whether they do it out of their existing, uh, sites, um, or dealerships. Um, or it’s some new location, it’s TBD, but I think what Tesla, so if if Tesla is doing everything. In-house then I think they already have the infrastructure in place. They don’t necessarily have to launch it, they already have it, but once they start selling it to third parties, um, it’s becomes interesting because Tesla can also sell them fleet management software and I think it’ll be great opportunity. ’cause I’m pretty sure Tesla already has it. They’re just gonna license it to other people. Um, and now you come to Tesla, let’s say you have few million dollars you wanna deploy. You wanna launch something in your, in your city. You come to Tesla and you say, Hey, I want to make money with Tesla. And Tesla says, great. Here are the Cybercabs. Here is the fleet management software. Here’s the network, here’s the contract, and then, you know, you kind of get it all. And then you go and register that with the state. You clear your insurance, you choose the site, you buy this whole thing. Tesla can even plug you with chargers. So you can buy those chargers and you’re good to go. And now you’re making money in the Tesla ecosystem. And, uh, Tesla’s happy. You are happy. Everybody’s happy.

Grayson Brulte: So that’s an interesting point. So let’s just, so that’s hypothetically say Tesla launches te, it’s called Tesla Fleet Management. Lyft, the company you mentioned, they acquired Flex Drive. They have their own in-house, which is doing well, that they’re managing the Waymo fleet in Nashville. Obviously Waymo’s using a series of partners, Uber, wherever they go, or using a series of partners. Could Tesla force the industry’s hand to bring fleet management in-house if it’s proven to be successful and highly scalable?

Sergey Litvinenko: It can go both ways. I think the advantage that Tesla has is that, um, uh, Tesla has been successful in many different businesses and, , if they want to do fleet management, they a hundred percent can do it. Other partners, it’s TBD. Um, because fleet management is, uh, heavy reliant on human operations, you need to have humans to manage those fleets. And, you know, it’s, it’s lot of little things that can get into the way. Uh, it’s not like a pure business. So I, I, you know, it can go both ways. I, I do think that. Tesla, in my view, Tesla will likely let other people do fleet management and have other companies pop up like move than do it themselves. But we yet to see, um, I just, I just not sure. If they run the numbers and they see, Hey, we’re gonna be making that much from FSD, that fleet management might look like a rounding error in terms of the potential revenue that it can bring. So maybe it’ll come down to that. We’ll see.

Grayson Brulte: You never know. We will see a lot of things. Sir, it’s been insightful. In your opinion, sir, what is the future of insuring autonomous fleet operations? And I put on purpose plural because there’s a lot more than just the vehicle.

Sergey Litvinenko: the future is bright, I think, for autonomous fleet. Owner operators because insurance is gonna be solved and it’s not gonna be a big problem. And I know the people, at least that we’ve been in touch to, they’re really worried about this because they have no idea what to do. Um, and for the insurance companies or reinsurance companies that participate in the autonomy economy, the future is gonna be even brighter because of, of how great this asset class is gonna be. So. I am extremely bullish and, uh, I, I, I really believe that this is a unique opportunity, um, to, to pretty much protect a new emerging market, uh, which we haven’t really seen before. I think the types of losses and the types of performance that we’re gonna see, we haven’t seen that before, and that makes it even more exciting.

Grayson Brulte: It’s very exciting times. If you want to deploy a fleet of autonomous vehicles, Cybercabs, or you just curious and you have the bank to do it, call Sergey and the team at Koop, they will help you underwrite the future. The future is bright. The future autonomous. The future is Koop Sergey as always sir Thank you so much for coming on The Road to Autonomy and sharing your wonderful insights to insurance with us.

Sergey Litvinenko: Thank you, Grayson.