Luminar's Investor Day | Oct. 6, 2020
(bright upbeat music) - Two of the fundamental barriers for autonomy are 3D sensing and perception. How well these vehicles see and understand the world around them. They can only make safe decisions based on accurate real time 3D data.
Solving these challenges requires a lot more than capital and smart specialized engineers. It takes the audacity to build something entirely from the ground up, hardware as well as software to make autonomy possible, bringing out of R&D and actually put in production. (music continues) - Together with Luminar, we have developed LIDAR technology with unprecedented perception capability. That means that you can detect not only that there is a human being here but also where he or she is moving.
- LIDAR is an absolute requirement for a safe autonomous drive. - It's a great challenge and a great reward to be able to develop autonomous technology for the future. This vehicle really represents the most sophisticated automated driving platform TRI's built.
- You know that little thing, it looks like a siren on the top of the car going around in a circle. Does every LIDAR system have to have that? - No, definitely not ours. Ours actually doesn't spin.
- Doesn't spin? - No. (music continues) When you see Luminar on the roof of your car, you'll know that it's a vehicle that safely enables true autonomy. We've always known this is possible and now we're the ones to have solved it. And that's why we continue on this pursuit every day, inspired to make autonomous transportation truly safe and ubiquitous.
Thanks everyone for joining. Certainly an exciting time. I'm Austin Russell, founder and CEO of Luminar hosting today at one of our LIDAR software testing as well as vehicle integration facilities in Palo Alto, California As for any company, entering the public markets is a landmark milestone that caps our incredible journey so far and brings us into a whole next chapter. With that, we really look forward to the opportunity to be able to share more details about the company, technology, product, commercial integration, and strategy with myself and some of the key members of our leadership team. Our vision is to make autonomous transportation safe and ubiquitous. And at Luminar, we're all incredibly passionate and committed to be able to make that happen.
The ultimate goal of this industry has always been around safety. And with that said, it's crazy to put even into perspective, just that we do lose 1.3 million lives every year out on the road from vehicle related collisions and deaths. And these are the things that can be preventable from technology.
That's the opportunity to be able to prevent. We don't have to have fully autonomous technology everywhere all the time to be able to make that happen but this will have just an incredibly massive impact on society. These are the things that we need to fulfill and these are the things that we can do today. That's what we're here to solve.
All right, let's go for some history and then we can dive a little bit deeper into the technology. So I founded the company about eight years ago with a goal to be able to build a new type of LIDAR sensing system for the autonomous vehicle space. And there's a number of very stringent requirements that are needed to build a system to be able to safely enable autonomy and be able to see it through into production in the real world.
So knew there was no way to be able to do this and meet the performance safety requirements, seeing out to 250 meters in the distance for all types of objects, with great resolution, much less also in a very cost effective device using off the shelf parts. And that's why we had to start from scratch, building our own components, our laser, receivers, scanning mechanism, processing electronics to be able to have something that can meet this specification. So with that vision architecture, started to bring on a couple hundred highly specialized team members and engineers to be able to build out the various components and systems in this architecture, to be able to see this happen in the real world.
Over the course of the first five years, we remained in stealth mode. We actually ended up acquiring a couple of companies along the way, including Black Forest Engineering, a chip design company and based out of Colorado Springs and Open Photonics which actually brought on a co-founder, Jason, into this, that you'll meet later. And with that, that's where we showed off to the world what was possible in 2017 with this breakthrough level performance. That was when we launched with four key commercial partners and over the next couple of years spent a little bit more time specifically on the commercialization, industrialization, maturity side of things, expanding from four partners to now over 50 companies that we're working with today and going through the various iterations of the technology.
We're actually well into multiple generations of our own ship designs, of our own architectures and all coming back to the same fundamental technology and principle, but continuing to iterate on it to a point of now, we can put it into series production. Further driving customer growth and accelerating programs is a software stack that we've successfully developed on top of our LIDAR. If the LIDAR is the eyes of the autonomous car, this is the brain, and this allows it to be able to autonomously understand what's going on around it in the environment and be able to safely navigate accordingly.
This year, excited to be able to get out there with our Iris product for series production. The Holy grail of the autonomous industry has always been to be able to take it out of R&D and put it into series production. We've been able to do exactly that with our landmark deal with Volvo, to be able to put our LIDAR and software and the next generation of consumer vehicles in series production starting in 2022. The fundamental reason for why we're here, comes down to the technology and what we've been able to build. We took a very different approach, whereas most companies in the autonomous vehicle space have started from the software side of it and worked their way to try to figure out what hardware can accommodate.
And we knew that the hardware and the sensing systems out there were not enough in terms of performance and safety to be able to ultimately solve this problem and enable autonomy to make its way into the real world. So we've been largely focused on the passenger vehicle side as well as the long haul trucking side of the equation here, really leveraging this existing multi-trillion dollar a year industry to be able to see our technology become ubiquitous and be deployed in the near term. Specifically, we're largely focusing on highway autonomy use cases for driver out of the loop functionality in those environments because they're more constrained than urban environments. That's one of the key areas that we see actually being able to be successfully enabled in the relative near-term.
We think there's a lot of promise over the long-term for level four or five urban autonomy which is largely where the vast majority of the companies in this autonomous domain have focused, but we see that operating in complex urban environments will still take a long time to successfully train systems to be able to handle all those types of edge cases. If you take a look at the broader sets and levels of autonomy, you can really separate out things into two discrete areas, assisted driving and autonomous driving, very different things. So assisted driving, the drivers in the loop constantly paying attention, ready to take over the wheel at any given moment, eyes on the road.
This is a reminiscent of the Tesla autopilot systems of this world and relevant other systems with other automakers that require constant driver attention and may follow a couple lanes on the road ahead. With autonomy, and by the time you go into that domain, driver no longer has to be constantly paying attention. You can start using your phone, working in a laptop, watch a movie, that kind of thing during that time and know that you can be safer than a human otherwise would have been. In addition to the highway autonomy focus that we have, we are also able to enable something that's often overlooked which is what we call proactive safety. In the areas where we're not autonomous, in suburban and urban environments.
In addition to, of course, when you do manually drive on highways, if you decide to do so then proactive safety will be able to help prevent forward collisions by actively taking control over the braking systems and steering wheel to be able to get you out of hairy situations altogether. It starts with the data, garbage in garbage out, and when it comes to safety, it's not acceptable. So you have to have a level of reliability that's been unprecedented before you have to be able to accurately detect all of these different types of so called edge cases to be able to accurately and safely be able to drive, and I'd like to show a few examples of what we mean by these kinds of edge cases and what, and how our LIDAR and our software solves exactly that problem. All right, so what you'll see here in these examples is raw 3D data coming from our LIDAR with different colors representing different distances, as it goes out with perception, the detection of those objects layered on top of that represented by bounding boxes. This is actually during a data collection run on the 280 freeway at night. You'll be able to see zooming way out in the distance, 250 meters ahead, a stalled black car out on the road.
We get seven points on that object, which is really a lot in this context. And 250 meters really represents just seven and a half seconds ahead at those speeds. So it's important to be able to see that full distance so in these kinds of situations you'd ultimately be able to come to a safe stop. If you zoom into 75 meters ahead, just a few seconds ahead we could actually be able to clearly make out what's going on, you can see the person, the car, the tire on the road. It's the person actually swapping out their tire that was pulled onto the shoulder.
And even at 25 meters, you can clearly make out even by eye what those different objects are and goes. Those are the kinds of things that we're able to now see with this per level of performance from a LIDAR standpoint, as well as our software for detecting and identifying the objects ahead. So in the next example, we'll show it LIDAR safety and performance is important even in lower speeds too, as well. If you take a look at this example, we have one of our data collection cars, driving around just after dusk. You could see some camera footage of the suburban neighborhood. And I don't know if you quite caught that, but there was a white orb that just rolled out into the screen, just in the middle of the road there.
And it's actually really hard to be able to detect particularly with just the camera, with everything going on you have the other street lights, you have some other lights in the scene, but this is where LIDAR is supposed to come in and help save the day. The challenge is even with these very expensive legacy LIDAR systems, if you zoom in on this data, you can actually only get a single point on that ball. And that's not nearly enough to be able to accurately detect something. It's usually an autonomous car would just see that as noise and be able to drive right through it.
With Luminar, it's a completely different story. That exact same frame, we can clearly make out what's going on with the ball, and of course, what's to come, a girl chasing after it. That's what makes all the difference. You can see our software identify the different objects and you can actually even see the scene play out live with a girl running out into the street and the car coming to a safe stop just in time. Those are the examples of edge cases that are absolutely needed to be able to successfully enable any level of autonomy to be able to clearly recognize and see.
And of course are still extremely helpful for assisted driving systems. For many, the question of how and when autonomy will successfully transition R&D and be put into the real world has been outstanding for some time. Just a few years ago, the predominant assumption was that urban ride hailing robo-taxis would be the defacto way autonomy is realized in the real world, in a city near you by 2020. Of course, that didn't exactly happen but did at the time resulted in an explosion of companies focused on R&D in this domain.
The rationale among these autonomous vehicle companies was, one, that sensing systems it costs tens of thousands of dollars ultimately needed to be amortized over the cost of the vehicle in 24/7 ride hailing operation hence the large roof racks of LIDAR systems and other things on these vehicles. Two, there was no LIDAR that could ever meet the long range, high resolution performance requirements that are ultimately needed for high speed highway autonomy driving as compared to the low speed urban driving that required shorter range. And then number three, is that autonomously navigating city environment would be a straightforward problem was the assumption despite the massive complexity associated with edge cases in those urban environments. Today, none of these assumptions have proven true, the first two reasons in large part because of us. And now that we have a high performance system that can see it long range and be a low enough cost to be able to put onto production consumer vehicles, it's something that was unfathomable just a handful of years ago. With that, we're the only autonomous vehicle company to be focusing in on this market.
And at the same time, we're powering nearly every major autonomous trucking company out there. The economies of scale leveraging the passenger vehicle and trucking markets are also enabling this to be used for assisted driving use cases like the proactive safety system that we talked about to be able to prevent forward collisions and accidents ultimately altogether. And this is how we can see the technology standardized throughout the larger industry and make as big of a difference ushering in the whole next generation of vehicle technologies and safety systems. Launching this bold vision forward, we entered into a landmark deal with Volvo for the first automotive series production deal for autonomy in the industry. Our hardware and software is integrated into Volvo's next generation consumer vehicle platform to enable these highway autonomy and proactive safety features scheduled to start production in 2022. Historically, Volvo has been the industry leader when it comes to safety and they've invented everything from the three point safety belt back in the day, and introduced most modern new types of active safety technologies that have paved the way for next generation vehicle system safety.
So with Luminar, expect it to be no different. We get to leverage the exact same product that we're building for the Volvo vehicles across the rest of the industry for other OEMs, for both passenger vehicles as well as trucks, in addition to the software too. That's really important to ultimately have a clear path towards widespread adoption in series production among multiple global automakers and over the longterm, standardization throughout the industry as with other safety technologies. Eight years ago, we completely re-imagined LIDAR technology building something entirely from the ground up.
Since then, we've successfully delivered on it and our LIDAR is the only one in existence that can meet the stringent performance, safety and economic requirements to be able to see this through into series production, to take autonomy out of R&D and bring it into the real world. The seamless integration of our hardware and software together ultimately enables a turnkey autonomous solution that accelerates the ability for OEMs to deliver autonomy in series production scale. We're not just a major force in the LIDAR space but also in the autonomous and auto industry at large. No other company has successfully built the LIDAR sensing foundation, much less the software. That's also required to be able to see this technology through, into the broader industry, in series production and that's what's made all the difference. As we've been making the transition from a technology development company to now global provider of autonomous systems to major OEMs, there's been a big shifting focus towards execution and that's something where it's been driving a major focus of mine to be able to build out a really strong team of leadership, the executor's that can be able to see this vision through end to end, and that's what we built out here.
Really look forward for you guys to have an opportunity here to hear from some of those members, it all started with the technology, and with that, I'd like to hand it off to Jason in Orlando. - Welcome to Luminar Orlando. I'm Jason Echenholz, co-founder and chief technology officer.
I'm responsible for the technology vision and strategy for Luminar. Here in Orlando, we're gonna give you a peek under the hood to see the technology, architectures and innovations that enable our products to deliver the industry leading performance that you've seen and you'll see the rest of the day today. In Orlando, we have 250 of the 350 family members in the company, and you're gonna get to see a little bit about the R&D and the core R&D architectures that enable our systems, the engineering and the capabilities to vertically integrate and deliver an automotive grade product and the advanced manufacturing team that allows our systems to be deployed around the world.
As Austin mentioned, we started with a Model G back in 2017 and shipping those to our first four commercial partners. Here, we develop the core architectures technology and building blocks that would enable our industry leading performance. The lessons learned in deploying those first systems from the Model G were invaluable.
That core technology innovation that we put into this then laid the building blocks and the architectures for our second generation system, the Hydra. We were able to improve our receiver technology and our laser technology and move forward to our third generation systems in the Iris that we're deploying today. What's common in all of these systems is the groundbreaking and revolutionary single laser, single receiver architecture that allowed an economy of scale, cost performance and system robustness that was unheard of in the industry.
So all of this technology enabled what we're shipping to customers today. Let's head to the receiver lab to take a closer look at one of the key components that enables our groundbreaking performance. We're here in the receiver lab where we take the chips from our Colorado Springs location, and we put the receiver together with the photo detector and build our receive module. Fundamental to Luminar is the that we don't have to sacrifice performance with off the shelf components. We customize each of our sub system from the chip level up. We custom designed the lasers.
We custom design the receivers, the scan mechanisms, and the processing of electronics and bring them together to offer the highest level of performance. As you may know, we operated a completely different wavelength of light, 1550 nanometers. We do that in order to unlock the highest level performance. The fundamental technology innovations that enabled us to go to this wavelike was developed right here in this lab. Traditional thinking was that these INGAS arrays were expensive and costs tens of thousands of dollars. We use some fundamental core technology of in gas, indium, gallium, arsenide, very rare material that typically is very, very expensive.
And people ruled out the 1550 wavelength because of that. We're able to take a very small piece of INGAS, a small flex, smaller than a grain of sand, and we combine that with our silicon receiver chip. We get the best of both worlds. We own lock the performance required to enable our LIDAR systems and the economies of scale, where we have a chip that only costs around $3. With all these technology innovations, you have to patent the heck out of this.
We have over 87 patents already in our extensive portfolio. In fact, it's twice as large as our top five competitors combined. That's it from Orlando, let's hand this off to Aaron in Detroit. - Welcome to Luminar Detroit. My name is Aaron Jefferson, VP of product, and I bring to Luminar over 20 years of automotive experience, delivering safety electronics, advanced driver assist systems, product leadership, and business growth. Speaking of growth, we're growing here in Detroit, the North American epicenter for the automotive industry.
I joined Luminar because I was excited about the vision and about the technology. We operate in the automotive industry that requires continuous development of performance and innovation, and to a market that is looking for technological advancements. And we Luminar bring that. There's a lot that goes from taking a product that is R&D and delivering that into serial production with the automotive grade quality and reliability demanded by the market. And if you look at this market and you look at all the key requirements needed to deliver the technology into the market, there are a lot of say sensors that have certain trade-offs and systems that have trade-offs and maybe they can do one thing well, and not the other thing or maybe a few things independently but we are the ones that can bring that technology altogether and enable all the requirements needed to deliver highway autonomy and proactive safety.
With this achievement, we now play a key role in bringing this pivotal technology to the market and making transportation safer. We do this in two ways, one by delivering on highway autonomy and two, by delivering on proactive safety. The nice thing about those two is the requirements for each are similar enough that we can focus in on a key technology and deliver that key technology enables our customers to unlock the capability and deliver into both.
First, let's take a look at highway autonomy, and we believe that initial autonomy application is on the highway and provides the most consumer value for the foreseeable future. Now, you might ask yourself, why haven't people delivered this before? Many have tried and there's a reason they haven't because all around technology hasn't been available to unlock and deliver that capability. However, with our technology, you can unlock the full capability of a highway autonomy system.
The reason that the sensing today hasn't met the need is because you need the range, you need the resolution and you need the perception performance to be able to really understand the scene and the environment and behave appropriately and safely in that environment. It's the reason we've won these landmark arrangements on the passenger vehicle side. It's also the reason why we are involved in nearly every major autonomous trucking activity and development activity on the market, delivering long haul, automated trucking on the highway. We foresee that the passenger vehicle market is still the market driver and the industry and we expect highway automated functions to grow at a CAGAR of 40% from now until 2030.
The automotive industry is also trending in this way, in terms of highway autonomy focusing on hands-off and eyes off operation. Now let's move over to safety, and our focus proactive safety. Recent data suggests that there's still 1 million lives lost annually due to automotive accidents. Today's ADAS systems really aren't designed to eliminate accidents. They're really designed to mitigate or lessen the severity of accidents.
For proactive safety, our focus is to eliminate accidents. Our LIDAR is capable of unlocking and enabling the full capability of safety at higher speeds in weather, and low light and can have the perception capability to detect cyclists, pedestrians, vehicles, children, and the most complicated environmental conditions. Many companies have developed LIDAR, but none of these companies have developed LIDAR to address the real market needs for LIDAR in the industry. And none have delivered upon the software required, and our software is fundamental to our system. I'll now pass it off to Christoph at Palo Alto. Take it from here.
- Welcome to Palo Alto. My name is Christoph Schroeder, and I'm VP of software at Luminar. Prior to joining Luminar, I led the software development team at Mercedes that developed urban autonomy as well as brought radar sensors into production at Bosch. Here in Palo Alto, we have 100 team members.
Most of them are software engineers and work on our perception stack. Beyond Palo Alto, our software engineering team is located in Orlando as well as Munich in Germany. Luminar's strategy was from the very beginning to build a hardware and software solution that combines both. The software team here in Palo Alto has been founded four years ago with the first software developers and has done a lot of research and development work at the beginning.
Software is extremely important to us as a company. With software, we're able to add additional functionalities and additional value to our product. We have the industry leading lighter sensor with the best performance. Here in Palo Alto, we work on adding the next level of capabilities and value to it.
To us a full stack software solution contains many different components. It starts with the enabling highway LIDAR technology that actually sees things. It contains things like the compute unit to process things as well as the entire software stack.
You need to understand deeply and exactly how the sensor works, operates, how for example our scan pair, and can be set to leverage that capability to build something on top of it that is much, much better than what you would be able to do if you just use the point cloud as it is. That's why our team internally is able to leverage those capabilities and build a perception stack that is much, much better, much, much more robust and much, much stronger than anyone else could actually do. When enabling autonomy, the key thing that you need to think about is to solve for the corner cases, the edge case, the thing that doesn't happen, unless it happens once in your life, you don't have anyone who supervises the car.
What does it mean for the car? It means for the car you cannot rely on anyone to be there in case something goes wrong in order to solve for all of those use cases, you need a technology that gives you the range, the resolution and the robustness to all the different environmental conditions to actually do the task in all cases. Camera technologies work well in a lot of cases and they work really well for some ADAS use cases but there are a lot of situations in which you as a human being, can't even see something. So how should a camera see it? It just not possible. The key foundation is gonna be the LIDAR. We are focused on solving two key tasks.
We want to enable the proactive safety as well as highway autonomy. In order to enable those use cases, range and resolution are key. Detecting small objects really, really far out is a key task, it's something that only can be done when you have range and resolution at the same time. You need to have the resolution to detect all those objects and distinguish them from each other, classify them and give the decision making software as much information as possible to actually make the decision.
Or what we focus on right now is taking all of the technology that we developed and put it into serious production vehicles. Good example for that is actually someone like Volvo who will take our LIDAR sensor, our perception technology, deployed it in their cars and actually put it on the road by 2022. In order to enable functions like proactive safety, as well as highway autonomy. - Hi, I'm Scott Faris, I'm the chief business officer at Luminar technologies.
I've got over three decades of experience and scaling optical component technology companies in both public and private companies. We're currently in our advanced manufacturing facility where we do both the manufacturing of our hydro platform as well as the pilot level work for our new Iris platform. Since the beginning, Luminar is focused on working with companies that believe in the future of mobility and autonomous mobility, as much as Luminar does. We focused on four companies initially with our first generation technology. Those four companies formed the foundation for some of our deepest relationships that we continue to have today. Working with companies like Volvo and Toyota research Institute, we're able to take that foundational first-generation product and really build the capability for the next set of customers.
We really wanted to main focused on working with the world's largest automotive manufacturers that were committed to seeing this through to volume production. There's a significant difference between development programs and automotive grade series production programs. Automotive series production programs really are the Holy grail of the industry.
Luminar has continued to invest in the industrialization of our product. This has allowed us to grow from our initial core four customers to over 50 customers on a global basis. Those 50 customers represent the vast majority of global OEMs. Additionally, those 50 customers can be broken into three key segments, including passenger vehicles, commercial trucking, as well as robo-taxis.
Collectively that group represents over 75% of a total available market. Today, many of our OEM partners have mature highway autonomy programs with expected launch dates between 2022 and 2025. With Volvo's productions expected to start in 2022, Luminar is extremely well positioned to leverage both the capital investments we've made in our infrastructure, as well as the industrialization of the LIDAR sensor itself to help our other OEM partners scale their autonomy programs on a global basis. Luminar is also working in the commercial trucking space. We currently are working with a significant majority of the global OEM, autonomous trucking application partners.
Luminar's technology and the ability to see at high resolution and extremely long range is particularly important for commercial trucking because the ability to see small objects as well as fast objects, such as motorcycles weaving through traffic, it's important to be able to assure an autonomous vehicle and particularly autonomous truck can operate safely at highway speeds. Short range LIDAR solutions offered by the vast majority of LIDAR manufacturers, quite frankly, aren't adequate to be able to operate in this type of environment. Once these commercial trucking applications are in production, it's gonna make a significant difference.
Autonomy is a true economic enabler for the logistics market. In addition, the benefits of proactive safety that we're able to realize in the passenger vehicle market also equally applies to the commercial trucking market. The other segment that Luminar has been focused on is the robo-taxi market. However, one of the limitations of the robo-taxi market because of the sensors that they've historically used has really limited them to low speed environments. Luminar's technology and the ability to operate vehicles in high-speed complex environments is really the key to unlocking these robo-taxi market opportunities.
At the end of the day, the biggest cost sensitivities and performance demands really are being driven by the passenger market as well as the commercial trucking market. Okay, so we've talked about the markets we're focused on, but now let's talk about where this is all really headed. We have 10 of our commercial partners that are deploying hydrant and advanced development application. These advanced development programs give us a significant competitive advantage positioning us to ultimately convert them in to series production awards. By 2025, we expect the passenger vehicle market to contribute the vast majority of Luminar's revenues with commercial trucking adding additional percentages over time. Taking a step back and looking at the true impact of the Luminar production program when we expect that over 1 million vehicles will be using the Iris sensor technology and really leveraging the foundational work that we did with the hydro platform.
- Hi, welcome back to Luminar Orlando. I'm Jason Wojack. I'm responsible for sensor development here at Luminar.
We're now in one of our production test facilities this is where our engineering teams are working on early sample builds of our next generation sensor, Iris. I'm really excited to finally share more about this news that we announced this morning. This is our first Iris and it's gonna be shipped to our partner Volvo this week.
Now we're gonna take a look at what it's taken to kind of reach this milestone and take you through the process. What's unique about Luminar and our LIDAR solution is that we have a single product to meet entirely what the industry needs. In order to do this, we had to deliver breakthrough performance in the point cloud and it's quality, but we're also gonna meet the cost that the industry needs to scale. We leverage the same core technology in our previous generations, but we've refined it for size cost, and power, and also to meet automotive qualified series production design. Iris has a unique design to get the best performance out of the sensor, we want it to stay at the top of the vehicle.
So that gives us the best point of view, to get the best performance. Because of the slim form factor, we were able to work with Volvo over the past year and their designers to create a very unique slim line integration into the roof line. This is kind of a seminal moment for the car industry and its design. As we open up autonomy, it's gonna act as an iconic design for the future and what initially established LIDAR and the integration of it.
So when somebody looks at the car now, they're gonna recognize that this is a car that has autonomy built into it. We've reached a significant milestone. We're at what is called the B sample phase for our Iris product.
What this means is that we've gotten into engineering validation testing and we're ready to deliver to Volvo. It's a significant milestone because it proves out all of our technology that we had in our other generations and shows that it's still capable in a smaller, cost effective, highly mass producible product. In order to get to this stage, we had to develop our in house capability.
This didn't really exist in the industry before. So we hired the key talent in order to make this happen. So we have an advanced manufacturing team that has developed the processes in-house in a way such that we can transfer it to contract manufacturers and collaborated with them on exactly what the process would be so that we can scale this up.
We're proving that process out now, we're developing the blueprint for that. We're gonna take that blueprint as we move from B sample into C sample, and eventually transfer that over to a contract manufacturer. As we transition through our B series production and into our C series, that's where we're gonna develop our production tooling, we're gonna start to lay down the blueprint with our contract manufacturers where they're gonna start to replicate lines. We're gonna move from thousands of units to tens of thousands of units, and then easily to hundreds of thousands of units with our contract manufacturing partners around the world. That's it from Orlando today, and I'll hand it back to Tom in Palo Alto.
- Hi, I'm Tom Fenemore. I'm the chief financial officer here at Luminar. I've been a finance leader in the automotive industry for the last 20 years. First running the global automotive practice at Goldman Sachs and then most recently at Jeffery's. We have a very exciting growth plan here at Luminar and let me walk you through it.
Our rate of growth continues to accelerate. We have over 50 partners today using our technology including seven of the top 10 auto makers. We've already been awarded two series production programs. Let me demonstrate to you the earnings power potential of our business model. In 2030, we estimate that the total addressable market for our products is over $150 billion. If we capture less than 4% of that market, we can make over 5 billion of revenue in two and a half billion dollars of EBITDA.
We are not your typical automotive supplier. We work with the engineering teams to put our cutting edge technology on their vehicles, because of this, we have amazing visibility into what their plans are to launch programs with highway autonomy or proactive safety features in it. As many as 10 of them are working on such programs with a startup production in the 2023 to 2025 timeframe. We have a clear path to execute on this revenue growth driven by series production contract wins. First, we've already been awarded two series production programs. Second, we are actively working with eight of our existing customers to convert those relationships in this series production wins over the next 24 months.
Only roughly half of those additional eight wins are incorporated in our financial forecast. To give you a sense of our success so far at the end of this year, we expect the order book for series production wins to be approximately $1 billion. And to demonstrate to you the exponential growth power of our business model. In 2025, we forecast that this order book will increase by over a factor of 10 to over $10 billion.
Today, our revenue comes primarily from two sources first individual unit sales of our LIDAR hardware to customers, primarily for test and development purposes as well as for the development fleets. Second is primarily from NRES associated with the development and eventual launch for these series production contracts. Starting in 2022, as we enter commercial production almost all our revenue will come from sales to series production programs and be categorized into three buckets. First, scenarios where we sell only the hardware sensor unit second, a solution incorporating our hardware plus software that enables proactive safety functionality.
And then the final bucket would be our hardware plus the software solution that will enable highway autonomy. Our business model is very scalable and it has very low capital intensity. This allows us to grow our margins, profitability, cash flows and returns at a very rapid pace as our revenue growth accelerates. There were three underlying factors that drive the scalability of our business in the capital light nature.
First, the same underlying hardware and software will be able to sold to other customers with minimal design R&D and capital changes. So as we continued to sell more and more sensors, we don't need to invest a lot more in R&D and capital. Second, as mentioned earlier, we are deploying a contract manufacturing approach. And then finally, as we grow our revenue and unit sales significantly we're able to amortize our fixed R&D, SGNA and other costs over a larger volume base and we're able to get significant purchasing power from our larger economies of scale.
To expand on this last point in more detail, our bill of material or BOM is expected to be approximately $500 per unit once we enter our first full year of commercial production. As we gain economies of scale, our goal is to lower that BOM to less than a hundred dollars per unit. This rapid decline in our BOM enables us to lower the price of our hardware to drive a rapid increase in adoption of our technology including standardization without sacrificing our margins. This will dramatically increase the safety of the vehicles not only save a lot of lives, but save a lot of time with the increased proactive safety in highway autonomy. Now, I'd like to turn it back over to Austin Russell, our CEO and founder for some concluding remarks. - Hey everyone.
All right, well, thanks again for taking the time and thank you to a leadership and Tom here with that as well, a big thank you to the media team for pulling off this production I know it was a huge lift. Certainly an exciting day, now out there with the announcement that we are meeting and delivering this first Iris series production unit. If there was any doubt about what we were able to do and what we could really pull off in this timeframe and deliver against this is the the key milestone.
Excited to be able to make that happen this week. With that, I think it would be great to get started on the Q&A side of things. I'm actually here joined with, with Tom as well as Michael Bero, our senior director of strategic finance and IR, I'll be leading some of the Q&A. This is the first time we've really done a Q&A kind of a broader public forum here.
So it'd be great to get this kicked off. Yeah, thanks Michael. - Many thanks Austin. For those of you online, feel free to type your questions into the chat box on the left hand side and we'll go ahead and address them as they come. Why don't we go ahead and open this up. First question, do you think there'll be multiple LIDAR winners in the autonomous car space or will it be a winner take all type environment? - It's a good question.
I think when it comes to the different verticals that you have to address, when there's passenger vehicle autonomy, commercial trucks, robo-taxis, and the adjacent markets I think particularly in the adjacent markets, there's going to be a lot of LIDAR diversity and as there already has been historically and there's definitely huge opportunities. We'll see actually a lot of companies, I think more and more pivoting into that domain, just given how hard it is to build something that can meet the spec needed to solve autonomous vehicles and specifically the long range sensing capabilities needed for highway autonomy. So I think with this, I mean, if you're able to execute to the degree that we think we can I think this could very well be a winner take all for that specific market and domain.
- Excellent, this next one's for you, Tom. Can you walk us through what the typical sales cycle is for an OEM when evaluating these sorts of technologies? - Sure, the typical sales cycle for an OEM is usually two to three years in advance of them actually starting production. They will select the prior suppliers for their vehicles and the parts that will go on them. We're a little bit unique because we're actually developing brand new technology and functionality for these vehicles.
And so prior to that, two to three year in advance warning that you typically get to be a supplier on a vehicle, you have to go through a development program for new technologies. The timing on that can vary by each of the individual customers, for certain OEMs it could take an additional two to three years for some OEMs, it can be shorter than that. Right now, we're working on at least eight development programs where our technology is actually on our customer vehicles and they're kind of developing those programs to eventually go in this series production contracts. As I mentioned earlier, at least eight of those that we're working on now, we expect to convert into series production programs over the next 24 months. And in our financial forecast that we shared with investors we have only about half of those or four incorporated into our financial model.
- Excellent. This one's for you, Austin. How do you compare your technology against Velodyne? - So, it's a good question. I think really it kind of as we spoke to earlier it really just comes down to the core tech of what we built out was entirely from the ground up making all of our own components, lasers, receivers, scanning mechanisms, processing electronics that are needed to be able to see these key performance specifications. And at the same time, really be able to see this product through industry's production in an auto grade capacity at the cost that's really necessary to have this industry take off.
And it's that trade off that you can also try and build a really high performance sensing system using a whole array of components, but it ends up either being way too expensive and still not performing enough, or you can try and build something that's really cost effective but then it won't be nearly enough performance to try to enable an autonomous specific application. I think ultimately there could be other LIDAR related systems for assisted driving related applications that maybe have some diversity there but as it relates to this, we are the only LIDAR sensing system that actually meets the core OEM specs that are needed to see this through with autonomy into series production. And then of course the software side as well, this is very unique to Luminar in terms of the deep hardware software integration that we built from the ground up. And that's what differentiates us as we kind of have taken things to the next level at the autonomous vehicle level beyond just the lighter components. So, going from a component company albeit high value components that are going in, but now a systems level company. - Fantastic. And sticking with that topic,
what are your advantages relative to say camera based systems similar to those Tesla? - Yeah, so it's a good question that really the whole point of LIDAR is that it gives you that true 3D data in the environment. Cameras are really good at getting some decent resolution around 2D data and images but this, you don't have to guess what's out there in that three dimensional plane. You don't have to try and extract and figure out what's there because when it comes to autonomy, you have to have very, very reliable detection. You can't miss things. Normally a 99% detection rate of something may be okay for a lot of applications for autonomous vehicles you can't miss one out of every a hundred people which it's actually not even quite there yet in many of those cases. So, you have to have ground truth understanding of the environment, you have to have 10 nines where the reliability that's what the LIDAR gives you in three dimensions.
And it's not just any LIDAR by the way, too, as well. You really have to have that level of performance and data fidelity to make the most out of it. That's where you have to have that camera like resolution for the LIDAR which we're actually able to deliver for the first time here and still be able to get that true 3D depth. Now, again, it's still complimentary to existing camera systems, we're not here to try and compete with those so to say, but this does kind of serve as the ground-truth center for these vehicles and programs and companies that we're working with. - Excellent. For this next topic,
how are you thinking about defensibility? How are you thinking about your IP portfolio and how do you protect the uniqueness of your technology? - Yeah, that's a good question. When it comes to defensibility I think this is actually something that is really solid for us. And it's a good question.
A lot of people will ask, "Okay, well, what's preventing this type of technology from just becoming commoditized over time." Like let's say, okay, you guys got the best stuff now how are you gonna continue to build this value and maintain this technological advantage? And I think the reason I think a lot of people ask when, particularly when it comes to for the hardware part of the equation which Luminar has historically been associated with is that a lot of hardware technologies do ultimately become commoditized over time and either lose margin market share, et cetera for what it may have. But part of the whole point is, is some companies actually are able to leverage that strong IP to their advantage.
You take a look at the mobilize and NVIDIA's of this world. There's no knockoff of those types of companies. The there's really high IP that goes into it and high value ultimately enabling high margins over long periods of time. And that's where really Luminar fits in, in terms of categorically.
So we actually have a largest IP portfolio as it relates to these sensing systems that of anyone in the industry. I think it's actually more than even the top five other related LIDAR, R&D efforts or companies combined. And the reason why that is, is just because of the fact that we really do have this core tech developed from the ground up. We're not using off the shelf commodity parts. That's what's allowed us to do this. Again, we build our own ships, we built our own receiver systems, laser systems that go into this and have also in parallel, locked up the supply chain for a lot of the key and relevant systems that actually go into this, even making key acquisitions along the way.
We mentioned earlier acquiring, the chip design company in Colorado Spring that was high specialized to one very specific component of the design and something that we've iterated on with custom chips for multiple components now. So, that's what's all led us to be in this hugely advantageous technical position but at the end of the day you need to see the commercial adoption too, as well. This is the automotive industry and getting into series production specifically. It's one of those very high barrier to entry, but equivalently very high barrier to exit type arrangements. And that's why we sought so important over the past couple of years to start getting designed into these autonomous development fleets with an eye toward series production, working directly with these OEMs, to be able to see these programs through.
That's why it's important to get embedded into that, be integrated into the stack and then it becomes more a matter of when, rather than if that program will successfully materialize into series production and with Luminar as the key system powering it. So that's how we're highly defensible in a way that's differentiated from pretty much any other company in this space. - Great, and we have a few questions coming in on the commercial opportunity.
So aside from Volvo, or in addition to Volvo, can you walk us through the relationship that you maintain with the other 50 commercial partners? - Yeah, so really as it relates those, I think Tom mentioned that it can kind of separate it into different stages. There's validation stage, where we have, we work with companies where they initially (mumbles) our sensing systems. Basically do a host of tests, validate and do the diligence, we meet the specs that we say we do all those of course in all paths with flying colors and I don't think we've ever really lost a LIDAR shootout so to say, as it relates to performance and safety or even ultimate longterm economics but when it comes down to the capabilities here and how these programs progress it generally starts into that phase. We progress it into these kind of advanced development contracts. That's where we start working closely with these companies programs fleets.
We have to be selective about that, we probably have realistically more opportunities than we can even successfully dedicate resources to to be able to take on and see through, two series production at this stage or why we have to remain focused. We started to actually working with 10 of these programs now at the advanced development stage. And, I think eight of which are very promising towards, with an eye toward series production.
And that's always been the Holy grail of someone to be able to achieve that milestone of getting into series production. Of course, that's what we're now doing in the Volvo case it's been public that we're out there and it just an to what one other OEM in that context. But as it relates to this, there's absolutely gonna be a lot of stuff ahead on that front. And I think as people progress through the phases that's gonna be important to being able to realize this not just on the passenger vehicle side, but also on the commercial trucking side as well.
So I think it makes sense to keep an eye on that. - This one's for either of you, how should we think about the $150 billion TAM in 2030? - So when we look at our TAM as Michael mentioned, we view it as about $150 billion in 2030. As we mentioned earlier for systems that deploy our proactive safety system we view the content per vehicle at about a thousand dollars per vehicle for that system.
When we look at systems before that deploy our highway autonomy solution both on the hardware and software side, we've view that as about a $2,500 content per vehicle solution there. And so when we look at the forecast for potential LZR to L2 proactive safety opportunities in 2030, plus the highway autonomy there that is a key contributor toward Tam. Once you start getting into robo-taxis and commercial vehicles there, we believe there's a much stronger value proposition for our products on those vehicles which could drive potentially higher pricing on the passenger vehicle side.
And then the other thing here is that robo-taxis and commercial vehicles will need to deploy three to four sensors and sometimes in certain cases, five per vehicle. And so when you kind of look at all that and combine everything there we get to about $150 billion Tam in 2030. And then as you go out to 2040, when you see a much higher penetration for those levels of autonomy, as well as that's when we expect that robo-taxis next decade to become more prevalent we are seeing our Tam increasing significantly to North of $500 billion. - Okay, and then how do you foresee the LIDAR software landscape evolving over time and will software play a larger part in the ecosystem? When will this be either be recurring software licenses or one time? - Yeah, it's a good question. And as it relates to that, there is a distinction here.
I think the reality is, is that when it comes to the software side of the equation, most people have really tried to separate out this autonomous stack and different discrete components. We have a lot of the larger autonomous vehicle companies the Waymo's and Cruises and everything of this world that have focused in on the robo-taxi domain. They've been developing software for that, trying to work with other LIDAR related companies in this domain to be able to get something off the ground and then ultimately shifting towards the right solution for series production when is at that stage. The thing is though, is that when it comes to passenger vehicle applications, and actually getting something into series production, there really hasn't been any software that's successfully developed to be able to really realize this, see this through as it relates to processing of the LIDAR data and systems.
This is a very different type of approach when we're focused on highway autonomy specifically as well, these proactive safety systems then trying to develop a robo-taxi. We have to actually develop something that can be deployed in a series production vehicle, auto grade, level and quality of code and a performance capabilities that your life can really depend on. So that's what makes all the difference there and there's no question that is absolutely gonna be playing at a substantially increased role to accelerate adoption of this throughout the industry. I mean, just a LIDAR alone is actually not super useful to the vast majority of OEMs. In fact, you can't really even see it commercialized until you have the software that can successfully enable it and even enable the applications that it needs to be able to go into. So that's why we saw it as key to developing the software side.
We're actually at a point now where even internally to the team, we've had an inflection point where you actually have just as many software engineers as we do hardware engineers and that's really what's driving the vision of this forward. It's what we said, it's a huge portion of kind of use of proceeds for some high ROI opportunities to be able to accelerate the programs and also even further expand our Tam. Tom mentioned the total size of content value that we can address on these vehicles.
That's how we get to those numbers is because it's both a combination of hardware and software. Again, we will ultimately, and we do sell this all a cart. So it's not our way or the highway, if people have different other components of the systems, we absolutely work with multiple automakers to be able to integrate that.
But it is absolutely a critical part of the story and you will not see a successful industry without that. - Okay, on the topic of data strategy, will Luminar be able to harvest this proprietary data much like Tesla does across its entire fleet? - Yeah, also a great question. Absolutely, this is definitely... this is one of the things that Tesla got right.
And when it comes to getting hundreds of thousands of vehicles out on the road that can effectively go back, collect data for you, so you can continue to improve a system. It makes all the difference for that given set of applications. Now, of course there's a significant difference with collecting camera data for an assistant driving application, following a couple of lanes on a road and having the level of ground-truth LIDAR data that you need for the equivalent to some of these other autonomous vehicle fleets that have been out there collecting data with this. I think the key distinction here though, is that, whereas some of the largest autonomous vehicle fleets out on the road have had in the hundreds of vehicles out there going and collecting data in every areas like, Arizona or Phoenix or Pittsburgh or wherever it may be we actually have a kind of breakthrough opportunity by getting into series production to have hundreds, not hundreds, but hundreds of thousands of vehicles out there collecting data at the global scale necessary to see this ultimately realized everywhere. And that's a really important point because, I mean, if you were to try and deploy hundreds of thousands of vehicles on a test and development fleet to try to equip your own cars driving around with a hundred thousand dollars type of roof rack setups I mean, it would cost you tens of billions of dollars, heck you might as well just buy a car company at that point.
So, as a result, that's why we actually instead of paying to put these data collection vehicles on the road, we actually get paid to be able to do that. And that's a huge and key advantage that we see ultimately leveraging on the software side they kind of builds that hugely defensible layer on the software, beyond the hardware too, as well. - Now, while we've already done the comparison of Luminar versus Velodyne, can you just walk us through some of the more nuanced differences between 905 nanometer and 1500? - Yep, absolutely. So 1550 nanometers, it's a longer wavelength of light.
It's actually known as an eye-safe wavelength. People in the LIDAR industry have actually like there's nothing fundamentally new about using a 1550 nanometer wavelength of light for LIDAR people recognize it as could say the superior wavelength for some time, if you could find a way to be able to make it work. And the real challenge with 1550, as opposed to 905 which uses a more commodity, silicon receiver components as well as existing laser supply chains is that there are no readily available off the shelf components at 1550 to build a viable system at the performance level and much less level of economics that are needed. The advantage of the 1550, just to clarify though, is that you can get dramatically higher pulse energy and peak power is associated with this the LIDAR side of it, to be able to see the necessary long ranges and resolution with only a small fraction of number on it. We've obviously been proud of what we've been able to build from using just kind of a single laser receiver based architecture, as opposed to using a huge array of components and that's in large part just because we've been able to be so efficient at this 1550 nanometer wavelength with a small fraction of the components.
But again, we had to build out all of our own supply chain, build out all this from the ground up. I have this cost of in gas that Jason was mentioning earlier go from tens of thousands of dollars for a wafer to now just using a fleck of it for three bucks powered with our chip and in volume. And that's the kind of differentiation and distinction that's allowed us to succeed and also contributes to that whole commoditization question. This is fundamentally differentiated hardware that took us like moving really fast, five, six years to even develop the true first iteration of ultimately nearly a couple hundred million dollars that we had to be able to see that through with some of the most highly specialized top engineers in the world, in their respective fields working on this. So that's what we saw through, but it's here today. It works and now we get to ship it starting into series production with having our Iris sensing system all come together and start shipping this week.
- Great, and let's stick with that particular topic, Tom, this one's for you, around the cost of the system, one of the key inputs for making a decision and certainly for the OEMs, aside from volume, what are the factors are contributing to our ability to really bring down that cost over time? - Sure, as Austin mentioned earlier, we have very few components particularly on the laser and receiver side in our product, particularly for the Irish unit that we just shipped the first one today and will ultimately go into the series of production vehicles. And so when we look at our BOM or bill of material for our first full year of commercial production we expect that to be around $500 per unit. More importantly, as we go from making 10,000 units per year, tens of thousands of units per year to making a hundred thousands of units per year, and ultimately hopefully millions of units per year, we expect our BOM to decline from roughly $500 to approximately a hundred dollars per unit. Now, most of that is from economies of scale and a portion of that is gonna come from some engineering changes that we make to our next generation after Iris. And so it's really what's driving that is relatively few components that we have on that.
And then because these are custom designed components that Austin and the team effectively designed and built and incorporating into Iris, as we start buying them in bulk, we're gonna be able to drive our costs down to those targets. - Okay, this is a two-parter for you, Tom, do you think the current downtrend in the stock market associated with specs is associated with broad based sentiment? And then the second part of this, can you lay out the timeline to close this merger? - Sure, let me take the second question first. So, what we've said publicly is we expect to close our merger with the Gores group in Gores Metropoulus this quarter, so by the end of the year. I think the exact timing will be dictated upon our SEC review process. But right now what we said publicly is we expect to do at this quarter or by the end of the year.
With the first, so I used to be a former investment banker and so during that time I would kind of share