Written by David E. Zoia (Wards Intelligence) / Posted at 5/17/21
Q&A: AImotive CEO Talks Sony Car, SiliconChips and New Market Opportunities
Reposted with permission from Wards Intelligence.
Disruption in the battery-electric-vehicle market could be the ticket for Hungary-based automated-driving developer AImotive, its top executive says. In an interview with Wards Intelligence, CEO Laszlo Kishonti says consortiums being formed by Sony and others outside the traditional automotive landscape to design, equip and build BEVs – many expected with advanced features – may present a big opportunity for companies such as AImotive. Sony, for instance, has chosen AImotive to provide its aiDrive autonomous-driving software stack for its Vision S car. Sony showed off the Vision S concept at CES in 2020, and this year announced a prototype was being tested on public roads.
The Japanese electronics giant has said it doesn’t intend to build a self-branded BEV for sale, contending the Vision S is simply a technology testbed, though speculation continues the vehicle will enter production by 2024. In addition to AImotive, Sony is working on the project with a handful of key companies, including Magna Steyr (manufacturing), Vodaphone (5G connectivity), HERE (mapping) and Elektrobit Automotive (software), as well as Valeo, ZF, NVIDIA, Bosch and Qualcomm.
Kishonti stops short of suggesting Sony will indeed enter the market, but he points to the electronic giant’s collaborative approach as emblematic of the industry’s new, more collaborative direction when it comes to BEVs. Electronics companies such as Apple, Huawei, Xiaomi and a handful of BEV startups look to be following game plans similar to Sony’s, meaning as a developer of a complete plug-and-play advanced driver assistance system platform, AImotive could have companies knocking at its door.
“There is one trend that I see in the market (and) this strongly correlates with Sony,” Kishonti says, alluding to its cooperative approach to BEV development. “Sony (has) a very good brand (and) very good relationship with their customers. I think they just saw that by using their strengths in the market, combined with strengths of others, they can build an interesting product.”
That push into the market by startups also will pressure existing OEMs to match technology on their upcoming vehicles, and many of those automakers also may look for outside expertise to develop competitive products quickly, Kishonti says.
In the interview, the AImotive CEO talks about the impact of COVID, the need for automakers to control the supply and development of microchips to ensure the autonomous-driving future, the potential for going public through a (SPAC) offering and where he sees the company heading over the next five years. Here are highlights of that discussion:
Wards Intelligence: Has the COVID pandemic been a huge setback for AImotive?
Kishonti: We expected about 50% growth in our revenues from 2019 to 2020. At the end, we reached about 10%, or almost flat. I think that was because many of the customers changed into cash-saving mode, so they postponed everything (that wasn’t) necessary. In our case, many of the projects are targeting future developments, so everything slowed down. But I think this year they are coming back, so nothing is lost forever.
Wards Intelligence: Where are you seeing the business coming from for aiDrive?
Kishonti: We see that many (new BEV) companies through SPACs now have a few billion dollars. A few billion dollars is a lot of money, but it’s not super big money in the automotive space. (So) they (now) have a clear mandate: They need to make cars and get on the road as soon as possible. There is no time to build everything on their own. There is no time to build their own factories. There is no time to build their own automotive software stack. Indeed, a large piece of the (aiDrive) customers will be – or I expect to be – the new (BEV) players who want to get into the market. But the second important potential customer base is the not-so-big OEMs who are either late in the game or they honestly admit that they don’t have the resources to compete.
Wards Intelligence: When will we see a vehicle on the road that is using aiDrive?
Kishonti: I think it’s two years from now, but hopefully less.
Wards Intelligence: Would that system be for Level 2-plus ADAS technology or all the way to Level 3?
Kishonti: Many of these vendors are planning to build cars with sensor setups that can theoretically be Level 3. In some cases, like (automated valet) parking, (it would be) Level 4. But at the beginning at least, they will release features (that) are market standard and stable and safe. They plan to (provide) those additional features (later) as over-the-air software updates.
Wards Intelligence: aiDrive initially was primarily a vision-based technology. Are you starting to incorporate lidar?
Kishonti: We can show the customer three different versions of the same neural network. One just gets camera and GPS. Another version adds the radar. A third version adds the lidar. Each are better somewhat, because additional sensor modality gives you a better sense of the world, more stability, more coherence. But I think there will be a huge market for the vision-based (technology), because it’s...good enough for (use cases that)…still require the driver to be overseeing the system. Lidars at the moment are still rather expensive. Vision plus radar is the market standard right now.
Wards Intelligence: Valet parking appears to be a couple of years further down the road. Is that a technology issue or is it market acceptability?
Kishonti: It’s a network issue. We have cars (that can self-park). But it’s a totally different experience if you go to (an unfamiliar) parking garage and you need to map with your brain and your eyes yourself first time. Where is the entry? Where is the easiest (place) to find the empty spot? It sometimes takes 20 minutes just to figure out. If we don’t have maps (of parking structures), then the cars will take 20 minutes to find their way (too). That’s not very pleasant, so we are cooperating with companies like HERE. They are building a series of maps for the parking garages. Our algorithms are prepared to use those, but the coverage of these maps is probably not wide enough (yet).
Wards Intelligence: Expectations for autonomy were quite high three years ago, and now reality has set in a little bit. What do you think is likely over the next few years?
Kishonti: (Silicon chips) is where the big change will happen. In the past, OEMs told the chip companies, “We need this chip for that price (and) functions.” They built something and five years later, the chip was used in a real car. Now, this is simply too slow. If you check your iPhone or Android phone, every year there is a new much, much better chip released with a better phone. Now, some of the car companies (that are developing in-house software expertise) are recognizing…they also need to have better control over the chips, because this will be crucial. When you have the software, but you cannot release your car or function because you don’t have access to the right chip, that will be a painful experience.
Currently, those chips (that) lack supply are not super fancy chips in most of the cases. But the demand for fancy, high-sophistication automotive chips will be one of the trends for the next three years, because that is the base of integrating all the different functions in the car.
To build better products, integration is key. If (automakers) are serious about software, and they want to spend billions of dollars on software development, then they should have already recognized that they need to have a chip strategy. Not just to get supply, but to be able to increase the features in sync with the software and be able to differentiate from their competition. Because otherwise, there will be the Android car and the Apple car, and all the OEMs will be just South Asian white-label production companies.
Wards Intelligence: So, we have to get to that software-defined architecture on board vehicles to get autonomous technology to move forward?
Wards Intelligence: Where do you see AImotive in three years, five years?
Kishonti: We always had revenues since day one. But I want to make sure that our revenue will be not 30%, but 90% production-based. This is the big step we need to make, (to have) most of our development related to real cars and real use cases of actual consumers or drivers. That’s what I want. I think we have a good chance that if we will be still independent company, we will be the leading independent technology and software provider in Europe.
Wards Intelligence: Do you see an IPO at some point? Or maybe a SPAC approach to going public?
Kishonti: There is no venture-capital-backed company in the world today which doesn’t maximize these current opportunities. We see companies which are not more mature than us do successful SPACs. The reason we are thinking and still not doing it is because I was in the investment space beforehand. I was managing about $3 billion as an investor.
I saw that for public companies, the market tide can turn very easily. (Going public is) an opportunity to get a few hundred million dollars now, (and) then we would be free from finding new investors for years. But it is also a risk that if all of my employees spend half of their time watching the very risky share prices, that’s also a thing we need to think about.
Also, many of the SPAC companies are overpromising at the moment. There will be big busts, and do we really want to be part of that? That’s my question.
As an investor myself, I wouldn’t like to invest into SPACs. Would I be a SPAC? Of course, if there is a good opportunity, we would do it. But at the moment, there is lots of liquidity on the market, and things rise with the flow of liquidity, which makes me a little bit uneasy about this.
Wards Intelligence: Do you see your company as a Tier 1 supplier or a Tier 2?
Kishonti: We think of ourselves already as a Tier 1 supplier. (Working) with classical Tier 1s, the relationship is not always easy (for us), because they see what we do is (provide) intelligence to the hardware, and this is exactly where they want to be. Selling an automated driving strategic software stack to Tier 1s, I think most of the time we hit the wall, because they want to do the same. They know that that’s where the value is generated.