Sumit Sharma
Analyst · Ladenburg Thalmann
It's a great question, and it’s from my favorite topic, so I'll try to keep it concise instead of talking too much about it. So, I think if you think about these sensors, a lot of them, redundancy is going to be dictated by regulation. A sensor's capability, what will happen over a long period of time with actual miles driven on enough cars for consumers or testing, regulatory bodies like NHTSA, Euro NCAP, they will decide what's required. But as you can imagine, for every -- the big business is OEM. Obviously, a single OEM, let's say, shipping 10 plus million vehicles compared to like some shipping only 0.5 million. Things are going to change, as we know, but they will decide, which is the most scalable product, long term, accommodate for the regulation requirements, but anybody would want a solution, regardless the regulation requirement, solution that's the most concise, that gives us the best opportunity possible to ship a product at a competitive price. Adoption price of increases the lower the number of sensor counts. So, I think everybody knows this. If you're just watching the news, cars have radars and ultrasonics and LiDARs, and multiple camera modules, but then there's another bigger proportion of it that is not talked about often, which is called ECU, where all of this is fused together, and that's where the decisions are made based on planning and maneuvering. So, if you think about some of the choices we've made, some of the hard problems that we chose to solve for multiple years before launching our product, obviously, range is important. That's a basic. You have to have a range, so certainly have that. But this high-resolution where something at 250 meters as tall as me probably even shorter than me, significantly shorter than me, you could identify that is a very compelling thing. So, resolution is a very important one. On top of that, cluster velocity, so you can predict how things are moving in relation to the car you're driving, the ego vehicle, that's very important, again, in the same sensor. This whole point of like 30 hertz are low latency, this is very important because camera modules that are in the cars right now, they operate there. So, sensor fusion becomes a lot simpler. Unlike if a sensor was at, let's say, 5 hertz, 10 hertz, 15 hertz, some people would say, oh, it's all the way to 120 hertz. Well, we can be there as well. But what we do know is sensor fusion requires a very, very simple computing to merge them, simple, relative speaking. But then you can start seeing pieces come together that make the entire system, including the computing and substitute required to deliver true L3, L4 features as regulations are written for them. So, in general, having a sensor LiDAR with such high-resolution at the latency, which is about the same as a camera module and a computer system that can actually fuse the sensor and perform all the analysis on the fly, that's the path to a scalable product in general. Now what else stays in there? What are the features out of there? It's regulation, of course, what OEMs want to offer, what their differentiation would be. So this is the data all the way back from my time at Google and other places. This is the problem that everybody has been looking to solve is how do you take the number of sensors required to achieve autonomous driving and advanced safety features, and this is the path. So again, this is what I believe, is what we believe, and you can see in the market right now, you see different people feeling different moving platform testing. And we're excited about having this product available for them to put it on their moving platform and explore this.