AI & DePIN is Building the Future of Mobility: A Conversation with DIMO’s CTO, Yevgeny Khessin
““DIMO’s goal when we started in 2020 was to fix all the problems that had occurred in the 10 years prior in the mobility space.””
Decentralized Storage Alliance's Ken Fromm had the opportunity recently to sit down with Yevgeny Khessin, CTO and founder of DIMO.
They dove deep into decentralized physical infrastructure (DePIN) and decentralized storage and how DIMO is powering the future of the automotive industry.
Welcome Yevgeny! Thank you for taking the time today. Please introduce yourself to our community.
Of course. I go by Yev for short. I was born in Ukraine and came [to the US] about 25 years ago. Ever since I was a kid, I have always loved cars and so, unsurprisingly, after college I joined the automotive industry. I have been working in the connected vehicle space – building integrations with EVs, mobile applications, and other automotive services for years all with the goal of making my favorite asset – the car – as smart as your phone. We started building DIMO over five years ago with the goal of bringing connected vehicles to the masses and allow people to build applications and services within an open ecosystem.
Starting a company is not an easy thing. What got you to the point where you wanted to take that jump?
This is actually not my first startup. I started a mobility consultancy firm in 2016 which we grew from two people to 30 – mostly working with Ford. Funny enough, there’s a correlation between this first startup and this second one. Back in 2016, there was a lot of buzz around mobility and everybody had this vision of making transportation more efficient by making the car smarter, but unfortunately we didn't quite achieve it.
DIMO’s goal when we started in 2020 was to fix all the problems that had occurred in the 10 years prior in the mobility space. What the industry lacks is an ecosystem which is an approach that is net beneficial to all – as opposed to being a competition of who's going to get more vendor-locked customers onto their API. We've been building DIMO for about four plus years now based on this premise.
Can you explain the DIMO sites – DIMO.co and DIMO.org?
DIMO.co is our consumer offering and it’s where our mobile app resides. We use that to build the platform, validate it, and provide users with value who don't have a connected vehicle today. There's also
DIMO.org which is an open source project/protocol. It primarily resides on GitHub today and uses the website as the front-end/landing site for the protocol itself.
Can you explain what DIMO is?
DIMO is really an ecosystem. At the core level, it's a protocol that creates and supports an ecosystem. We've built all the plumbing in order to be able to ingest and store data – and most importantly – with users' consent, share it to a third party. The protocol, itself, can handle any kind of IOT asset – it could be a doorbell, it could be a scooter – but we've been heavily focused on cars because it's the one that can have the biggest net benefit to society.
Can you explain what data you're collecting, how you're collecting it, and ultimately the problems you're solving?
The DIMO protocol is open source and one of its features is a module called the Oracles, which allows anybody to build a data integration from their data source to a DIMO storage node. The integrations that we've built in in the open source include the Tesla API which you can find on GitHub. It lets any Tesla driver upload their vehicle to DIMO and start storing their data.
Another one is an integration with an aftermarket device company. If you have a car that is older, you might not have any of the telematics readily available to you. We allow you to buy a device to ingest and store data from this aftermarket device.
As for the data we’re collecting, vehicles run a protocol called the CAN bus. You can think of it as being just like your home wifi network. A car has its own local network where there is somewhere between say 50 to 200 data points that are easily collectible from the car.
Most of it is things that you would expect, such as speed, engine RPMs, location, tire pressure, which gear you're in, the odometer all the way through, is the trunk open, and is there enough diesel exhaust fluid in the car. All of these data points get stored in the DIMO node. With the protocol and the ecosystem, we're enabling people to build applications on top of this individual and aggregate data.
Maybe you’re a fleet company that has 200 vehicles for deliveries or trucking, logistics, whatever it may be, and you care about vehicle health, you care about where your drivers are, and you care if your deliveries are going on time. All of this data can be used to build a fleet application.
The best example that we're working on today that I think explains it clearly is let's say you're signing up for insurance in the world of pay per mile – like Metro Mile for example. What you're really agreeing to is I am going to give the insurance company access to my harsh braking, harsh acceleration and top speed and my approximate location on the other side. What you're really agreeing to is pay the insurance company an X amount of USDC per month or per mile driven.
What DIMO enables is this single unit agreement, which is bi-directional flow of data and payment that is signed by the wallet of both the user and the insurance company. It's a beautiful single source of truth that either side can cancel at any moment in time. All of the other pieces of the web3 finance infrastructure to then be used to onboard money in or offload it out.
There's also a gamification side of things. We have a rewards engine that you can use to provide incentives to users. So for example, maybe you want to build an application that verifies that you went to a Speedway station or another gas station or dealership and provide you a kind of airdrop reward for going to this location. We have a protocol module for this. These are just three of the many use cases that are live today.
What makes DIMO different from the others in the space? How do you set yourself apart?
Every other connected telematics system that exists today is a closed ecosystem. To be honest, I wouldn't even call them ecosystems. I would say they are SaaS products with preselected applications. DIMO is built on an open ledger, which means it's accessible anywhere globally. The smart contracts and the vehicle identity is accessible to anybody and anyone could use this to build on top of. Because of this, we have SDKs that most connected systems do not have.
We also have privacy built in. Part of building a protocol that's focused on IOT data sharing is how do you really gather consent in a secure way? This is something that most companies have done in a very basic sense where maybe there's one button you click to share data, but it is really enforced at the front door only. DIMO’s ecosystem has privacy at the core based on the private key signatures and a ledger system that allows everyone to know that user's privacy will be preserved – whether it's the application developer or the user themselves.
And so you've created an open data marketplace where the data is owned by the car owner (or maybe it’s a truck owner or a fleet owner) and where people can let that data be used for various purposes. Can you explain where you see this going?
Interestingly enough, when people talk about data marketplaces, they usually talk about one-way value flow – with data going one way [and value being captured by the aggregator]. With DIMO and the privacy model, you can build a data relationship and a cashflow relationship heading both ways. It becomes bi-directional.
Let's take something that people don't like doing and that is servicing your vehicle. When you use DIMO to permission the diagnostics and create a relationship between the repair shop and your car, there's a couple of bi-directional data flows that are happening.
One is the mechanic wants to know the problem your car is having so they can help you fix it. They want to know what error codes are occurring and so that's the data heading to them. Now, obviously they would like to get paid when they resolve the problem. Because DIMO is on a ledger and we've set up this bidirectional relationship, you can actually use onchain payments to pay for this transaction. What's even more interesting is the DIMO protocol allows the mechanic to issue an attestation that they serviced your car and that the problem was resolved. This enables the ecosystem to really thrive because the next time the user shows up, possibly a different mechanic, different dealership, maybe they want to go sell their car, they're actually then able to permission all of this data to these parties.
The system builds on itself because next time I show up, I have more data than I had before. The dealership or the service center six years later can still access this data. And when I go to sell the car, I have a perfect history of my vehicle that I can use to price the vehicle better than the automotive marketplace can use to price it.
Do you sell devices and where's the data stored? How do you verify the data as being from the car? Let's go through some of the technical aspects and why again, what makes DIMO unique.
The uniqueness does not come out of devices themselves. Every device that is needed to collect vehicle data already exists today. By the way, our goal is to build an ecosystem. It is not to sell lots of devices. We provide a mobile application and a plug-in device that you can buy on Amazon. There is a data storage node that is hosted in AWS by a third party and it can be accessed via them and also via the protocol’s telemetry.
Anyone can host one of these nodes and as a DIMO user of the protocol, you can actually change in the smart contract where your data is going to be stored. There is a second node getting spun up in Europe as well. The system functions the same way, just that the data storage will be in a different location.
The way we verify the data is via a mix of attestations. For example, every DIMO device signs data with the key in a secure element. You can think of it as we've added a Trezor [wallet] to your car. All the data that's coming from the vehicle is signed via the wallet in the vehicle which is then used by the protocol to verify the data source. We call this particular attestation proof of vehicle.
Why is it important to have this form of trust and verification?
My answer here has actually changed over the years. When you have an open system, it opens up various gaming vectors, right? DIMO is an open system. Anybody can call the smart contracts. And as we move into this world of AI it means that AI can now generate infinite data.
You're already seeing how easy it is to generate a video of somebody saying something they have never said. The same thing is going to reach the IOT space where it will be possible to generate IOT data at scale. You're going to need to verify somehow that this data has actually been generated for real, especially for insurance purposes but for many other purposes as well.
The second part is this open approach. Every single IOT system that exists today is closed source. With DIMO, the smart contracts and the ledger are accessible to all. You can have an ecosystem of assets that interoperate, that people can freely build on, that people can build solutions for user problems in a way that could not be done before.
[With the old way,] you have to go to one automaker, one charging company, one smart home robot company at a time to integrate all these things. Now it can all be done in an ecosystem approach where things are built on each other and what one developer builds helps another. One data point can be used with five to six different applications.
How do web3 innovations like tokens and ledgers make this possible? Why can't you do this with just traditional cloud and web2 approaches?
I think it comes back to the problem space. A car today is built in Japan, shipped overseas to the US, leased to someone in Florida, driven and serviced and insured for five to six years, and then sold as a secondary secondary market to Europe where it once again gets serviced and used. So that's the reality of the world today, right? Enabling what I just described of the lifecycle tracking of creation of the vehicle – if it was done in a server that's only available in Japan – it wouldn't function by the time this car made it to US shores.
Once you're in the US, if there is not an open ledger that people can build on, you're going to be left with the world we have today, which is the only applications you can use, the only ecosystem available to you is the one that the car maker allows you to use. And then when the vehicle goes to Europe, again, all this data, all these connections, all these applications are no longer accessible to you because in a different region with different rules, different laws. But the ledger and the DIMO network are global, the privacy guarantees are global and the applications can be global because of it. And so there’s this unlock of using a distributed ledger in terms of the user value.
There's also the privacy and verification aspects. Being able to build a system where you can verify that a user shared their data with this application is something that's very important for regulators, for users, even for developers. Developers, for example, want to remove the liability of how the data was made available to them. They want to be able to say, “This is when the user gave me data and this is how I acquired it.”
This is hard to do with a centralized SaaS business because the only proof is coming from the inside, whereas with DIMO, all the transactions are user initiated. You can see the exact timestamp when they sign the transaction to share their data. You can see the timestamp when the vehicle was manufactured and when it was first signed into the network.
Talk about digital currency as a native primitive means in the context of mobility and transportation.
The finance side is very interesting. People have been trying to figure out how to do automated changing for gas, EV charging, and parking for the better part of 10 years now. But as far as I know, there's nothing I can buy in the market today that actually enables this.
The automotive world has proposed a few different specifications that were, to be honest, way too complicated to go to production. You can't issue a Visa credit card to a non-human, for example, which is a funny limitation of the web2 finance industry.
But the web3 finance stack enables this. It can give a vehicle a wallet and it can give the charging meter a wallet. Instead of this impossible world that exists today where an EV driver has to essentially install six applications and add a credit card in six places just to be able to charge their car, we can give a wallet to every car and to every charger. This can just be done automatically on web3 rails with no new Visa credit cards, no new complexity, no new hardware.
What kind of reaction are you getting from people in the industry? As you explained the vision, are people getting it?
We're in conversations with insurance companies and automakers to integrate more natively with them. It's something that is going to take a little bit of time to get to. If the moment occurs that DIMO is natively integrated in a million cars per year coming off the assembly line, then we’ll effectively have transportation and mobility on a ledger.
At the moment we're running POCs and pilots with these companies and so they're getting it, but it's going to be a bit of time until you see the mass scale that the finance industry is starting to see. It took, what, four years from about when DeFi summer first happened to seeing Stripe announcing that they're going to run their own L1 and having these beautiful onramps and offramps.
For us it's very similar. You're going to start seeing these automotive and mobility style companies integrating DIMO over the next three to four years, but it's going to take some time.
The old saying it's an overnight success that took five to 10 years.
I would say. I mean there will be successes before then, but the joke we always make internally is that DIMO will be replaced by teleportation. That's the moment that I would love to get to.
But if I could speak to one thing specifically, which is that for automakers, using smart ledgers for mobility is not a new concept. GM looked extensively at this, and Toyota's blockchain lab has been publishing white papers on this topic as well. But it's going to take time and it requires the village to build a new kind of ecosystem.
DIMO is really seeking to do something with an industry level of scale to it, which is to enable drivers to bring their data and share their data in a way that they control to build this future ecosystem of mobility applications focused around privacy and user control and putting the driver first.
Having data that's owned by the users, that's stored securely, that can be verified as such, and that can be used with permission for a variety of uses – that sounds like a really good vision. How do people get a hold of a DIMO device, where do they go?
If you want to join DIMO as a user and you're interested in getting vehicle analytics insights on your car, you can download our DIMO mobile app today. If you have a Tesla, you can connect just via the integration in the mobile application. If you have a car from other brands, you can purchase our device on Amazon and connect your car that way.
If you're a developer, we have our developer documentation and a developer console. It's an open source project and so we have over a hundred repositories that comprise the DIMO ecosystem. Feel free to go there and start contributing or even just star it to like and follow. I would say these are the main areas.
To do a POC or otherwise get in touch with us, you can join our Discord as well as email us at developers@dimo.org.
We really appreciate Yevgeny taking the time to speak with us at DSA, the Decentralized Storage Alliance.
Make sure to follow: @DIMO_Network
Visit their site: https://ai.dimo.co/
Developer site: https://www.dimo.org/
DIMO store: https://dimo.co/products/dimo-lte-r1
