Back to Blog

Cowen Mobility Disruption Conference

Melanie Stone |




Joseph Giordano: Hi, everyone. Good afternoon. Thanks for joining us. My name is Joe Giordano. I cover industrials, robotics and automation here at Cowen. I'm happy to have the management team from Cyngn with us today. We have Lior Tal, CEO; Don Alvarez, CFO; and Ben Landen, VP of Business Development. I can take you through a presentation. We'll have a little bit of time at the end for some Q&A.

So, Lior, do you want to take us away here?

Lior Tal: Sure. Thank you, Joe. I’ll jump right in. So, for anyone not familiar with the company, Cyngn is a software and data company. We develop self-driving, software intended specifically for industrial and commercial applications. We've been developing our system since 2016. And as we're getting beyond the research and development and taking the first steps into commercializing it, we took the company through an IPO on Nasdaq in Q4 of last year to raise the funds needed for the go-to-market.

I'm going to go into some of these in more detail, but just as a highlight the software, the product that we developed is proprietary. It was developed in- house, it is a full-stack self-driving system. Yes, it is purpose built for industrial and commercial applications. So, when we designed the system, we thought about the full spectrum of uses from logistics, and manufacturing, distribution, mines, ports, everything you can think of under the industrial umbrella.

We are going after a very attractive immediate market with the material handling focus, which I will talk about. The major value in focusing on these industrial applications is the massive cost of labor that this technology can replace, and the added value beyond the available labor that it can assist in increasing productivity. We are managed and have a very experienced and very diverse team that has built and scaled companies, both private and public. And we work in collaboration with a lot of companies around us in order to take the full solution of the vehicles and the software into the end customer's hands.

Lior Tal: Now, the reality is that all major industries rely on massive human workforces as the backbone of their operation. And these people that operate across all these industries do mostly repetitive, predictable physical work. It's the picking and dropping, it's the driving and repeating these workflows many times per day. As part of that backbone all industries have large fleets of industrial vehicles that currently can only be operated by human drivers.

The challenge is that people are not really good at this repetitive type of work that requires accuracy over many hours, day in, day out, What people are really good at is the creative part of the workflow: solving problems, coming up with ideas. And the problem is that when you have the discrepancy between the dependence on humans and the imperfect match of humans to this workflow, you start getting into problems that span from accidents that happen because people are distracted, because they are tired, shortage that come because of the fact, this is hard work and not too many people want to do it.

And people that actually do start taking on these roles they churn, and enterprises end up with lost opportunity and lost revenues coming from the fact there is a labor shortage. And of course, there is the whole massive cost of operating these large workforces. And if you bucket all of these together, you see that the opportunity here to solve the problem of the dependence on human labor inside these industries end up being hundreds of billions and over trillion dollars over the coming years.

Now automation, as a technology, is just the perfect solution to all of these problems because by deploying an autonomous driver on this vehicle, enterprises can shrink and expand their fleet depending on the demand. They can turn on more licenses during the holiday season or during high demand times, and they can scale it back when they don't need to, and they don't need to impact the workforce by doing that.

Our software is inherently safe because it sees 360 degrees around the vehicle. It calculates much faster than a human can, it’s never distracted, it's never tired. So, by design it is supposed to do the work it is used to do.

And of course, you can operate it around the clock as much as the power of these vehicles can support. You can operate these vehicles, means you can add a third shift when, for example, your human workforce can only operate one or two shifts at the same facility.

Now, in addition to the business and the productivity gains, there is a very major sustainability impact especially by the combination of electric vehicles and autonomous. These zero emission vehicles combined with much more effective software operating them means that you can move from the previous generation of combustion engines and also operate them with much less energy. And over time in large fleets in major enterprises, this really contributes into ESG objectives.

Lior Tal: Now, the first industry we are focusing on is defined as material handling. These are vehicles that move parts, or goods, or material around large facilities of different types. You can see a few of the examples of vehicles here that are used for material handling. This one application from the full spectrum of industrial uses is already a massive one in the sense of how much is spent on the cost of labor to drive these vehicles. And you can see this is above $100 billion per year just for this one industrial application.

Now, while eventually we believe that everything that moves will become autonomous, there is a sequence on what will happen first and what would take longer. And I think the strongest contrast is in comparison to the open road applications. And we're all following the Tesla Autopilot and everything that's happening with highway trucking and passenger vehicles and robotaxis in comparison to these open-road applications that keep dragging and are going to take still a long time to get to really large-scale commercial deployment. The work that we're doing with the industrial vehicles with the commercial industrial applications are simpler across each of the variables that build the problem.

These environments are simpler, they are more structured. The workflow is more predictable. The speeds are slower. And the whole environment is much less regulated, which means that you can take this technology and apply it into these applications and achieve commercialization at large scale, much faster and as early as this year.

Now, because of the way we designed our system and because of the way our software driver, the AI component of the system was designed, it can be extended all the way up to working eventually on the open streets. But like I said, there is a reason to start with simpler industrial applications, commercialize them, and scale them before moving on to more complicated ones. But as a system, we have the ability over time to for more complicated applications added. And as we see that the team is ready, and the industry is ready, start taking on more and more of these verticals and expanding our software into more of these applications.

Now, our long-term vision is to be the driver of choice for all these different industrial applications, whether these are mining trucks, or forklifts, or stockchasers inside facilities. And the competitive advantage that we offer the customer that would use the system and our technology is really being able to harness the efficiency, the safety and the productivity gains from automation in their operation. And that can be either the disruptor and the first mover in their industry or they’re one that chases a larger company that has already put their hands on automation and is really starting to harness the gains of the technology.

Now, the real breakthrough six years ago, when we started building the system was really scanning the industrial automation and the industrial machinery space and identifying the technologies that are used for automation. They are just not cut for the work that these vehicles need to do.

The robotic navigation systems get stuck a lot. They drive slow and a lot of it is because they don't have the intelligence to really understand like a human being, what is really happening around the vehicle, what the mission is, and really make the decisions that allow the vehicle to move around things and resolve events that happen in real time.

And the way we approach the problem is taking the robotaxi design of a self- driving system and harnessing it downward into these simpler industrial applications. And by that we're able not just to take advantage of the design of the system, the software components, but also the supply chain and the talent and everything that the large companies working on open road are investing into to enable their business.

So by that breakthrough of really taking the robotaxi approach into the industrial applications, we were really able to get into a generation of solution that is above anything that is out there today, and allow us to work really in collaboration with the existing vehicle manufacturers, the ones that are really the vehicles of choice of these large enterprises and install our software on their vehicles and do it either as a retrofit for an existing fleet or off the line with new vehicles coming up in the coming years.

Now, because our approach is to develop a software solution that is separate from any one specific application, we're able to then work with a customer that has heterogeneous fleets and works with multiple types of vehicles that do different tasks. And you see here an example of nine different vehicles that we deployed that software on over the past several years, that span from floor scrubbers to stockchasers to electric trucks and people movers.

And what's really different in our approach because we're focused on that software is that all of these different vehicles in their different environments and their different applications are using exactly the same software. And that is very leveraged from a customer perspective because it can then work with a single vendor and all these vehicles, the different vehicles in their environment can really benefit from the other vehicles around them. They communicate with them, make the whole thing safer and the whole deployment and maintenance cheaper.

And of course, the superior level of driving because the system was built in much more complicated environments, driven on open road, but always with the understanding that it has to be deployed into the industrial applications, it drives as a human would drive in these environments. And doesn't suffer from what the older technologies in the industrial automation are applied with.

Now, if you really go a layer underneath that, what really allows us to be able to offer these solutions and compete with what a human workforce is able to do today, is our product that's called the Enterprise Autonomy Suite, EAS. And EAS is built out of two parts that are more customer-facing: DriveMod, which is the software that is installed on the vehicle and completely autonomously operates the vehicle, and Cyngn Insight, which is the set of tools that the customer interacts with whether to manage the fleet or control it or benefit from the data that comes from these vehicles.

And beneath the waterline we have what we call Cyngn Evolve, which is the set of tools that allows our researchers and developers to continuously evolve the system, keep training it, keep improving it and make sure that it is future proof.

DriveMod, like I said before, is designed like a robotaxi brain. Like a robotaxi driver, it has the ability to use many types of different sensors using LiDAR as the primary one in order to see the world, understand which types of objects are around us, be able to differentiate between people and vehicles, and other objects that are unique to the industrial environment we're operating in. And then, because we know what these objects are at this point can really track them over time and start predicting their motion.

Now by understanding the scene and the environment around the vehicle, we can then make very complicated decisions in real time in reaction to the changing environment around us, which means that instead of just driving and reaching a point where the vehicle gets stuck in front of an object, we can preempt that by understanding that that object is not going to move, so we need to nudge or overtake it and then just continue driving. And all of that translates really into productivity and uptime of the driving fleet.

Once that decision has been made, it is then transferred into the adaptive control stack, which was designed to support many different types of vehicles and really learn how the vehicle operates, which allows us to deploy the software in different sizes and different categories of vehicle and repeat the whole cycle up to a hundred times per second. And this is part of what makes the whole system so safe and so flexible over time.

Insight was designed for unskilled labor. On the other side of the equation where the customer is able to take people that before maybe drove the forklift and train them to now support 20, 30 vehicles because they know the work, because they know the vehicle. And our tools are very intuitive and very visual, it's very easy to then migrate people that before drove vehicles to be site operators or to be fleet supervisors and really take their capability into the next level where really the human capability kicks in and augments what an automated fleet can do.

Evolve really works similar to how Tesla’s pipeline works, where you can harness all the data that comes on the vehicles, we can decide what to bring up into the backend, whether for new machine learning or AI training or for simulation. New versions of the software can be tested overnight driving many, many miles without even needing to move the vehicles. And when we're comfortable the new version of software is better than what we have in the field and we want to send an update in the press of a button, we can send it down to whatever part of the fleet we want to. And then when these customers start using the vehicles the next morning, they're already upgraded with a new version.

Now with this, I'm going to hand this over to Ben Landen, our VP of Business Development, to talk a bit about the plan to take this to market, the steps we made recently and how this really translates into the commercial part of our journey.

Ben Landen: Thanks, Lior. So, there's a wide array of different customers here because it really takes a very expansive ecosystem to deploy the complexity that is autonomous vehicle solutions. They range from technology providers that provide complimentary technology to service maintenance providers, but I really want to focus on the vehicle manufacturers. That is really our key partner, and what you see here are two vehicles from the Columbia Vehicle Group that traced the journey that we've had with them over the last 18 months or so.

Starting with the yellow Stockchaser that you see there, these are vehicles that can tow thousands of pounds or be loaded up on the bed directly. That yellow one was the first prototype that we did an in-house retrofit to prove out to our partner Columbia how their vehicles would operate autonomously. And then with several revisions along the way in between, what comes out on the end is that teal and white version, which includes our patented DriveMod sensor module on there essentially pre-calibrates and streamlines both the retrofit process and the end of line coming off the manufacturing line integration process. So, it maintains the ability to integrate both assets that are already out in the field and fresh ones coming off the line, which is exactly what we are working to do with Columbia and it's the type of engagement that we see being mirrored with other manufacturers down the road.

Let's go to the next slide. So going back to what Lior had mentioned, this is a bit of – a lot of that was ideas, strategic positioning of ours in terms of how a software defined system, how a robotaxi doing industrial vehicle types of jobs would work. This shows that the proof is in the puddings. So, this captures a few key highlights from the years that we spent building the technology stack, starting all the way back in 2016 when we started to develop these AV solutions on the roads of California in very complex mixed traffic environments. Then we went and were very selective about the pilots that we did at the time. Those ranged from light utility vehicles in the Philippines, where we were exposed to rains and changes in weather at a seaport and all the way up to five-meter-long shuttle buses that we deployed at the corporate headquarters of Loblaw, which was a 4,000-person crowded complicated campus. As you see, we were even blessed with snow operation there.

And then transitioning past that phase where we were building the technology. Now we step into the IPO and what that takes us to, which is commercialization, productization, and scale. So, these are just some of the milestones that we've had over the last few months that we're excited to share. We did a successful pilot deployment at Global Logistics and Fulfillment’s Las Vegas fulfillment center. I'm going to hit all of these briefly because as you see, there are links there's available collateral for this on the web ranging from press releases to videos. We did lean in further with Columbia on building autonomy ready Stockchasers within their fleet. So that means revamping all of the components that they're using, ensuring that the best possible version of a Columbia Stockchaser is coming off of the line. So that is ready for the integration of that DriveMod kit for which we filed a patent and also mirrors our ability to retrofit Columbia Stockchasers that are already out in the field with many, many Columbia existing customers.

We also announced our partnership with Greenland Technologies where we will be embarking on our forklift effort – forklift automation effort. Forklifts of course are one of the most ubiquitous material handling equipment pieces out there. So very exciting and we love the disruptive nature and EV focus that Greenland brings with sustainability and maneuverable vehicles, and we really see a great synergy there.

So ultimately what that enables us to do: the fact that we have led with the software means that our business model is a subscription-based model, effectively SaaS. That SaaS if you zoom out even farther to the fact that in the end of the day the end customers are signing up to use a vehicle on a subscription. We do fall within the robots-as-a-service model and we specifically are the software provider within that, and it becomes a robot when we couple it together with our hardware partners. And this is a rather substantial pie that's there. Lior showed some of the market numbers that we see available due to the cost of human workforce that goes into making these vehicles do the work that they do. And non-trivial volumes considering millions of these vehicles are out in the field with 10-year useful lifetimes looking for retrofits. And that on the order of 900,000 new vehicles are sold per year, just from the top 10 manufacturers.

And now let's go to, I'm sorry, you hear my daughter just woke up in the background there and is not happy about it. Now to, to my favorite slide that pulls a lot of this together. There are three main takeaways for me in this slide. One is the dimension that you see that is opened up and it's kind of hiding in the cartoons here from the fact that we are vehicle agnostic, that we built the technology to do that from the start. We've proven that on nine different vehicle form factors, and that enables us to expand within sites because practically every customer site out there is using multiple different types of vehicles. And that ability to address many different types of vehicles is fairly unique to how we've built the technology.

Another takeaway is that that really creates a defensible situation where once we've done a first vehicle within a site, once we've done a second vehicle within a site, once we've started to expand to other sites with a customer, it becomes kind of matter of fact that we would do the next vehicle – the next autonomous vehicle for that customer, because we get the sharing of the data that we've already collected. It's the same repositories, it's the same interfaces, the same training and the marginal cost for us becomes lower and lower whereas the barrier to entry to an ad hoc supplier of a vertically integrated automated solution, it becomes much higher for them to break in.

And then the third takeaway here is that really, this does underpin what we have been communicating since we geared up for IPO over a year ago. This is the plan of record in terms of how we see our scale growing over the coming years and how that ultimately traces back to the previous slide and the revenue model into when we do see ourselves stepping into really scaled revenues after the paid beta deployments that we're doing through these Phase 1 and Phase 2 timelines.

So, to wrap it all up, the key takeaways we hit on these, it's a big market, advanced autonomy this concept of a robotaxi, kind of big brain solving smaller problems is what we really see as an accelerant to getting into a sustainable and sustaining company. Key strategic partnerships, this is massive. We're all about partnering about leveraging the strengths, the dealer networks, the service, the incumbency of folks that are already out there and have been doing this a long time and serving as an additive function. And of course, we're very proud of the team that we have ranging from our board, from our executive team to everybody that we've got contributing to this vision of the company.

And with that we will open it up for questions.

Joe Giordano: Great. Thanks guys. If anyone in the audience, if you want to ask a question, just either put it into the dashboard and I'll see it, or you can just send me an email at Okay. So, I'll kick it off. I just wanted to clarify now, are your customers like the – you're talking about the OEM, the manufacturer of the machines. Are you also selling to like, if I own a facility, I could bring you into like retro all of my different vehicles there? I just want to make sure I understand who you're targeting.

Ben Landen: Yeah. So actually, what we consider our primary end customer is who you said at the end there: the site operator, whoever's running the machines, that's really the end customer. We're addressing them together with our OEM partners. That's how we provide that software-as-a-service plus a robot or a machine to achieve that automated vehicle for the customer.

And with that said, there are other revenue streams that we're vetting out that would come directly from our OEM partners or from other parties that either care about the data or build to a design – that we have designed where we then focus on the software. And there's some additional revenue streams opportunities there, but the end customer is the site operator, is the person – is the entity running the factory, running the fulfillment centers.

Joe Giordano: But you're able to put your software package on like an old machine?

Ben Landen: The simple answer is yes. The yellow Stockchaser that we showed that had all of the components on top was a vehicle that was built many, many years ago, that is made out of metal and old controllers and it's essentially – it becomes a sliding spectrum of the amount of effort that's required. But practically any machine with the right amount of elbow grease can – we can fit our technology to it. But the point is to streamline that and that's why we went through that effort with Columbia. It's why we showed them on a vehicle that wasn't great for that retrofit process, we showed them it was possible and then worked with them so that they're rolling better vehicles off the line so that we can streamline the process. And so that we can take those learnings and be more efficient for a customer who does come in and says, hey, I've got 50 of these in the field. What's it going to look like to start making some of them autonomous?

Joe Giordano: Okay. So, I was little confused, because it said, I think there was a slide that said it was all EVs and I was confused, I guess that's all new stuff rolling off a manufacturing line, I assume?

Ben Landen: We have a few different channels for that. Again, this is where partnership and working with folks like Columbia who have been selling vehicles into this space for 70 some odd years is a big strategic advantage for us. They've got existing dealer networks; they've got existing sales, service, maintenance networks. We most certainly tap into those. We ourselves have organic business development efforts to find new opportunities. And beyond account-based marketing, we do there is – there are bigger fish in this domain. There’s a long tail of smaller types of fulfillment centers, kind of mom & pop shop and that we would reach with traditional outbound marketing types of approaches. So, we have a few channels that all contribute to those ranging from incumbents account based and long-tail marketing.

Joe Giordano: So – okay. And you can also go through like the big Caterpillar type companies too, I guess as well?

Ben Landen: The simple answer is yes. Essentially, that – that's why we showed that expansive partner ecosystem. I only had time to talk about a couple of folks on there, but any one of those are viable channels to increase our business development efforts. For example, we haven't spoken publicly on the specifics, but ranging from IoT digital transformation, management consulting, 5G provider types of companies, those are partners that we look to strike within our partner ecosystem as well since it's a very natural stepping point into—well, the now how do I bring autonomous vehicles into the portfolio given that all of those technologies are being adopted?

Joe Giordano: So, I think it's interesting, like the vehicle agnostic approach and we can – you can make anything in your facility autonomous. I guess the counterpoint to that is, if I wanted to make – if I wanted an autonomous forklift, for example, like why go this route rather than get something from Seegrid, or Neckar or someone who makes autonomous forklifts specifically, like how – what's the advantage to this platform versus others that specialize on making the autonomy for a specific thing?

Lior Tal: I think the question is who are you, right? Because if you're Greenland and you manufacture the electric forklift and you want it to be autonomous, you don't go to your competitor and try to see if they'll give it the socket, right? So, the customers that end up wanting to choose a Columbia vehicle or a Greenland vehicle, because they like the vehicles, they like the network would also want to get the best driving software on it, right. And a lot of the incumbents, a lot of the vehicle manufacturers that ended up offer autonomy are really using the older solutions, the robotic navigation solutions. So, you have to compromise on the automation in order to get the full vertical solution.

Now, if you're a customer that has – a facility that has Stockchasers and forklifts and scrubbers, it means you need to work with three different vendors. Those solutions don't talk to each other; you need to replicate the cost of the whole thing, right? So, in a way, thinking about the software as an, almost an operating system for autonomy and being able to deploy it for whatever need you have in that facility is much more future proof and eventually gives you that productivity you got from people before.

Joe Giordano: And what's the typical deployment time?

Lior Tal: Off an existing solution, if you look at the Stockchasers that we piloted with Global Logistics, we need a day or two to learn the site. The team goes there with a data collection vehicle, collects the data, bring it back, and a few days later we know the facility. We can deploy the vehicles if they're available. If these vehicles need to be built by Columbia, there's a preorder process, but the deployment of an existing solution, a Stockchaser, for a different site of the same use case is not something long.

What will take longer is learning something new. For example, we deploy the Stockchaser, now we want to do a project with a mining truck, right. That means we need to learn that vehicle. Again, it's something that is not prohibitive and much easier than bringing the whole thing, right. But that will take a project around it and then scaling up that vehicle is going to be similar. It's going to be a matter of days here and there.

Joe Giordano: So, I know we're at the end of time. I just had one last one. How does your system work with all like how interoperable with other robotic systems? So, if I have AMRs moving around my facility from another vendor and I wanted to bring you in to do forklifts or whatever. How does that all work together?

Lior Tal: So, the vehicles themselves are self-sufficient. They don't need to talk to anyone in the field whether it's human or other. The backend inside the management system, the planning, the fleet management, it's an open system. We have APIs. We can integrate to other systems. We can have other systems talk to ours. So, it really depends on what the customer wants in terms of interoperability between their systems. But because these autonomous vehicles are really operated independent and are tasked by our central management system, it's fairly simple.

Joe Giordano: Great. I think we have to leave it there. We're a minute over, but Lior, Ben, Don, thank you very much for taking the time. It was very interesting, and we'll look forward to following the story.

Lior Tal: Thank you, Joe.

Ben Landen: Thank you.

Joe Giordano: Take care. 

Similar posts