Electric vehicle company Arrival wants to break the current auto manufacturing model. Instead of one giant factory and an assembly line, Arrival’s commercial electric vans, buses and cars are robotically built in small, regional microfactories, of which the company wants to open 31 by the end of 2024.
If you want to achieve something radically more efficient, you have to go deeper, into complex, high-level computational algorithms that are not normally used in consumer-facing products.
The London-based company, founded in 2015, joined the ranks of EV companies going public via SPAC, merging with blank-check company CIIG Merger Corp. in March. UPS has already ordered 10,000 of Arrival’s robotically engineered vans, and the company recently signed a deal with Uber to create purpose-built EVs for ride-hail drivers.
Arrival founder Denis Sverdlov has been at the intersection of technological advancement and societal change before. Born in the nation of Georgia, Sverdlov founded his first company at 22 selling IT consulting software to enterprise customers. Since then, he has built and exited multiple companies, most notably telecommunications operator Yota Group. Founded in 2007, the same year the iPhone came out, Yota eventually launched a 4G network across Russia, coupling it with an HTC smartphone that would facilitate the use of the network. Sverdlov sold the company in 2012 for $1.5 billion, did a brief stint as the Russian deputy communications and mass media minister, and then went on to start Arrival. Oh, and he founded electric autonomous car race Roborace in 2015, too, just ‘cause.
With the same sense that fast, mobile internet and big screens would change the telecoms industry back in 2007, Sverdlov began to see a perfect storm brewing in the EV world over the last decade. In 2015, he founded Arrival in anticipation of a switch to electric as well as advancements happening in material, research and development in the robotics industry. He predicts this will have an even bigger impact on the automotive industry than 4G had on telecommunications.
TechCrunch: Denis, your first company was a telecom operator and you were behind the creation of Roborace. Now you’re trying to change the way the auto industry makes cars with Arrival. Are you a serial entrepreneur who is already thinking about the next thing? Or are you pretty involved in this one?
Dennis Sverdlov: Yeah, I’m quite involved with Arrival, and I expect to see many new technologies and enablers come out of this journey. For example, if you take our robotic technologies, which we use for microfactories, you can easily see how that will be used in other industries, as well, so it’s not going to be used only for automotive.
But as a company, we need to focus on what we do today because we need to achieve a lot here, and I think it’s important to focus on that. But, yeah, I would also count myself as a serial entrepreneur because I’ve been doing this for more than 20 years.
How do you think that your past business decisions have informed your current strategy with Arrival?
Today, I believe that we live in a world where everything is technology. If your business doesn’t have a technology element, then it doesn’t have a future. And so what I know from all my previous businesses is that for radical change, you need to apply radical technologies, as well, which I would call “deep tech.” With full respect to companies like Spotify and Netflix, that level of technology we’re seeing is on the surface. If you want to achieve something radically more efficient, you have to go deeper, into complex, high-level computational algorithms that are not normally used in consumer-facing products.
More Transportation Founders
What’s an example of your deep tech that you would point to as something that’s a game-changer for Arrival?
I think our computer vision system, which recognizes car parts in factories. Usually, you can apply standard computer vision algorithms with machine learning to identify patterns, but you’re limited with the geometries of the parts you can bring to the factory, and with the fact that you need to train your neural networks for a long time. So if you design a new part, and you want your factory to recognize it, you need to spend a lot of resources to actually train your neural network to recognize that part in different conditions.
We reinvented a totally new vision algorithm using geometry features to identify our parts in the factory much quicker, no matter the environment, which will have a positive impact on production rates.
I can also point to our algorithm for multiagent optimization style of operations. What is a factory? It essentially consists of a lot of assets like robot arms, mobile robotics, some humans who are also part of the creations. You need to have very efficient algorithms to schedule all the activities in the way that you get the maximum efficiency there. And again, there is very complex math behind that, which you normally cannot just buy on the market because it just does not exist.
Speaking of what goes on in your factories, why do you think Arrival’s strategy, which is to build commercial EVs out of multiple microfactories, is better than building a single factory or working with a contract manufacturer? Why are you breaking the norm?
There are many, many layers that we can talk about. So first of all, how is the big factory defined? It’s defined by the current body of the vehicles. The majority of the vehicles we see right now use what is called a unibody, which is a metal body that makes up the structure and skin at the same time, and then all the components go inside. But in order to make a unibody, you need to do very heavy work using metal stamping, which necessitates a lot of pressure to get the shape you need, and for that, you need to have very expensive tools. So if you invest hundreds of millions in your tooling, of course, in order to be profitable you need to produce hundreds of thousands, millions even, out of these tools. And it doesn’t matter if you want to produce one vehicle per day or a thousand — you will need to have thousands of robots anyway because the process of creating a unibody car is defined by this technology.
Then you need to have thousands of people on the assembly line, putting the components inside the body manually because the body itself is already defined and there is no access to put them in robotically.
Another basic point is that when you buy components from Tier 1, like brakes or controllers or any other parts, you open the bonnet of any vehicle and see that you have so much empty space there. Why is that? Because when the Tier 1s were designing those parts, they were not thinking of how those parts will work together. They were solving their own problem, and that’s why you’re getting super strange shapes out of every component. Complex shapes mean that it’s not robotically friendly because robots cannot work with them. Our components are designed on a modular grid-based system, so it makes it easier for the robots to recognize the shapes and assemble them.
When I think about our business model, I like to use this analogy for it: It’s almost like switching from mainframes to the horizontally scalable architecture that we use in cloud computing. Nobody’s using mainframes anymore today because the architecture is limiting.
OK, so what’s the effect of making that change in production?
Our microfactories are twice as cheap as traditional factories. And for us, the major thing is that we said we are not using metal, metal stamping, welding or paint shops, and this was our condition since the beginning. This opened for us the opportunity to totally redefine the form.
This allows us to increase the level of robotic operations because we’re not limited by complex components and an existing architecture. And it gives us the flexibility to say that one factory can produce many models, which gives us an economical advantage.
The current factory works like this — you need at least three years to build a new factory, and if you want to run a new model, you need to stop production, refit the factory and then produce that model for up to 10 years. In our case, our microfactories allow us to produce 10 models from the same factory, and we can introduce a new model to the same factory without stopping production on the previous one. And this is a big deal for the industry.
My dream down the line, I will say as the father of this company, is to see the moment where we can do next-day delivery, rather than many months. This method is going to be an enabler of a new economical model that I believe is going to be commonly adopted. My vision is working forward to a model where producing vehicles in New York with microfactories is going to be cheaper than producing vehicles in China using the current method. I believe technology can enable new business models.
If we can switch gears quickly to talk about your funding strategy. Arrival was one of the most well-funded U.K. startups and secured a lot of really good partnerships early on. What was the logic behind pursuing a SPAC?
We invested around $600 million in the last five years in total. Compared to the other guys like Tesla and Rivian and Lucid, it’s a fraction of what they invested. What is important is that we are making four models of vehicles right now compared to other companies that did one model and invested 10 times more money. So how investment works in our company is very different to what is normally perceived as market standards.
Coming back to your question, of course we needed to have capital to scale. We want to build many microfactories; our business plan is to build 31 by the end of 2025. So for that, we need to have capital. We were in the stage of scaling, so it was the best time to go to market with a very strong product.
So why SPAC? Well, from my perspective, there’s not much of a big difference between a SPAC and a standard IPO process. The reality is, in terms of the amount of due diligence we needed to do and the amount of preparation within the company, I don’t see much difference in the investment of your time and efforts. What is important with this SPAC is that Peter Cuneo, the former CEO of Marvel, has an excellent background and lots of experience turning around four different businesses. He’s our chairman now and we see this not only as an opportunity to get funds for growth, but also to have access to those skills and knowledge.
OK, so choosing the right blank-check company, one that aligns with your company’s goals, is really important here. I have a prediction question for you about SPACs in general. EV companies and the technology that supports EVs made up about 68% of SPAC deals in 2020, and Arrival was one of them. Do you expect more SPACs this year? Do you think it can continue at this pace?
You have to look at it by the company, not define it by the SPAC. Arrival is unique compared to other companies on the market, so honestly, I don’t care how many EV companies come to the market via SPAC or not because I know we’re unique and powerful. We are the only company in the world that’s using grid-based architecture for components. I think there are smart investors who can understand this differentiation.
There’s a difference between quality of assets on the market, and you can have good assets and not so good. It doesn’t matter how you came to the market. What matters is the fundamentals of the company.
What do you make of predictions that all these SPACs are creating a bubble that’s sure to burst?
Look, I’m a technology entrepreneur, and I think for me it’s quite difficult to give comments about the financial markets in this way. But the advantages of SPAC for investors, for example, PIPE investors, is that you put the money in the company and it’s in an escrow account so it’s untouchable. You know your funds are there.
Then, when the founders of the SPAC make a proposal and say they want to merge with a particular company, you have a choice. You can say, you know what, I don’t like this deal, and you can take your money out without any losses. So this kind of frees a really unique opportunity for SPAC investors because it prepares them for an upside.
I realize it’s early on, but your company’s stock fell in its first few weeks as a publicly traded company. What is your tolerance for risk?
We as a company, of course, work in the interests of our shareholders; we want to do all the necessary actions to see the shares growing. And I believe that all our technologies and assets and the quality of management, the quality of the company is there fundamentally to support this growth. Short-term fluctuation of the shares doesn’t really make much impact on us because we have enough funds to make our vision real.
We lost like 30% in our shares, but within the company, nothing has changed from the moment we announced our deal. The business plan is still the same and the performance of our systems is the same, so fundamentally as a company, nothing has changed. It’s just the reaction of the market and many other moving parts, but I don’t see much of a connection to our fundamentals of the company. It’s not a reaction to our performance.
In the course of founding Arrival, what has surprised you along the way that you’ve learned from?
I cannot stop being astonished and surprised by how powerful technology is. You create something and you think that it’s a breakthrough, and you’re super happy. Then a few months later, you find that there are another few layers under that. Think about universal space. We’re constantly raising our heads up and we see this space and then we understand the concept of what is unlimited. In the same way, I see there is unlimited space in the technology field. You can go deeper and deeper. There is no limit.
Because we want to check in with you a year from now, I’d like to close by asking you where you think you’ll be in a year.
This is very much defined within our business plan. We plan to start production on our first microfactories this year, and that’s our main focus. So it’s a very clear answer: Our factories will start to produce vehicles.
This article was updated to correct the year Arrival intends to have 31 microfactories built.