Ouster, Inc. (OUST) Earnings Call Transcript & Summary

Welcome. I'm Blayne Curtis, semiconductor analyst at Barclays. Very happy to have with us today, Angus Pacala, CEO of Ouster. LiDAR has been a very interesting space, lots of newly public companies. Given that it's our autos conference, actually quite interesting. Ouster recently made an acquisition of Sense Photonics, bolstering its position in auto. So maybe, Angus, let's start it there, give us a little background on the deal, the rationale and your momentum in autos.

Yes, sure thing. So at the beginning of the year, we set out a plan to accelerate our global go-to-market strategy across all of our verticals, industrial, smart infrastructure, robotics and automotive; accelerate our hardware road map across our scanning and solid-state digital LiDAR sensors; and expand our software offerings. And Sense Photonics really hits at a couple of those initiatives and the acquisition of their team, their technology and their assets, which we closed this last quarter. So Sense Photonics is of all of the kind of private LiDAR companies that we've been tracking for years and years at this point. Sense Photonic stands out as the only one that's pursuing a similar digital LiDAR strategy to Ouster, a similar product strategy in a multimarket -- excuse me, a multi-sensor suite approach to automotive LiDAR, where you need to have various LiDAR sensors deployed on a vehicle with different kind of field of view and range capabilities, pricing and a similar go-to-market strategy in consumer ADAS automotive around focusing on winning serious production opportunities for hundreds of thousands and millions of vehicles and really going through a traditional RFI, RFP, RFQ process to hit those high-volume demand. So we saw that similarity of kind of mindset, team, technology earlier in the year and were able to get a deal done where they became a part of Ouster. And actually, with the acquisition we formed Ouster Automotive to kind of encapsulate all of that product strategy and go-to-market strategy in a single dedicated team to go out and win high-volume automotive applications in consumer ADAS.

It was bit of the pipeline, but maybe just a little quick bit on just what digital LiDAR is, VCSEL, SPAD architecture. Everybody has their own flavor of LiDAR, different lasers, different approach. What is differentiated about Ouster and now Sense?

Yes. So we -- Ouster has developed digital LiDAR technology, and that stands apart from analog LiDAR technology. And one of the home arts of analog technologies, there are many different ways to skin a cat in the analog world. And so if you survey the landscape of our competitor set, building analog LiDAR, you can build FMCW LiDAR and 1550 nanometer LiDAR and traditional that kind of scanning LiDAR systems, 905-nanometer LiDAR. But digital LiDAR is the concept of taking that complexity and integrating it onto a silicon chip and replacing all of those disparate analog approaches with one, an ultra-high performance and affordable and scalable silicon CMOS technology. And that's the strategy of Silicon Valley across many different verticals or many different technologies for the last 60-odd years, from CPUs to GPUs to telecom infrastructure to digital cameras. And so you really can't think about what Ouster is doing. Our goal is to integrate every last piece of complexity of a traditional LiDAR sensor onto a silicon chip and reduce the BOM count, reduce the part count and simplify the manufacturing by orders of magnitude compared to what's come before. And we've actually been able to do that. We've shipped multiple generations of our technology. If you look -- if you take apart Ouster's LiDAR sensor today, it's got approximately 100x fewer parts internal than a traditional analog LiDAR system from a competitor. And so, yes, build a LiDAR system onto a CMOS chip and get the benefits of affordability, scalability and actually exponential improvements in performance through improvements in the underlying semiconductor.

If you talk about your road map, you have spinning LiDARs, OS series, and we'll talk about maybe those end markets if we have time. But then you were working on a solid state for the auto market, the ES. Now you've acquired Sense's products. So maybe just walk us through where that fits in the room.

Yes. So one of the other unique things about Ouster is this idea that we have released product families that are composed of tens of different flavors of that product family, but they're all driven by one CMOS chip. So it's like putting M1 silicon in a Mac. You can put it in a Mac mini and a MacBook Pro and a MacBook Air. It's the same silicon powering a broad family of products. That's what Ouster is doing with our LiDAR sensors. So we have our L-series silicon that goes into all of our OS scanning sensors. And we're developing now this solid-state silicon that's going into our digital flash series of LiDARs. So before the acquisition of Sense Photonics, it was called the ES Series. Sense had their own version, their own kind of naming of products. What we've done is combine the best of the ES Series and the best of the Sense Photonics products into this newly released digital flash or DF Series products, all powered by one CMOS chip.

Right. And you mentioned that the chip the -- you just launched the L2X and you have an L3. That was the road map you kind of gave people. And maybe explain what that is? I mean, I guess it's the receiver with an ASIC. Do they have that same strategy and same silicon road map? And I guess, will you then ultimately, as you kind of combine this acquisition, still have kind of one receiver chip?

So yes, so we'll have a receiver -- okay, so what the receiver is or what this chip is? Is it's all of the pixels, and you can think of pixels like a digital camera, but they're more complex with more circuitry in them, in combination with all of the signal processing and command and control logic necessary to run a LiDAR system. So that's all on a single semiconductor dye, built in standard CMOS. And today, our L2X chip has over 100 million transistors on the chip in addition to all of the pixels. And we're going to continue to just pack more and more complexity from the broader system onto the chip and reduce system complexity by increasing chip complexity, because Moore's Law has allowed additional complexity on a chip to be essentially free year after year after year. And we're taking the same -- so that's the L Series processors. We released the L2X, the L3 will come, then the L4, L5, L6 and so on and so forth. We're going to do the exact same thing with our DF-Series products. We haven't named the chips there yet. The teams internally are still debating what they want to call them. I'm going to leave it up to the engineers. But we will have this product road map based around chip design cycles. In the same way that Apple every year, they release a new A-Series processor into the iPhone. The iPhone gets a refresh, but what's really refreshing is that new A-series processor in the iPhone driving new features and capability. Yes.

And maybe could we just change gears and talk about this pipeline because the Sense did have a pipeline when you bought them. So that definitely accelerated the timing there. There's a strategic deal. And then maybe if we could just talk about lot of peoples have development deals, right? Production wins are the key in this market. You've talked about having kind of 5 kind of potential production yields. Just walk us through that.

Yes. And so Sense Photonics, again, we've acquired them and we formed Ouster Automotive, which is really the broader sense team in combination with some of Ouster's people, devoted now to winning high-volume series production wins in consumer ADAS. And that stands apart from our other automotive verticals like robotaxis, robotrucking and shuttles and buses, which is another part of our business. Actually, we generated 40% of our revenue last quarter from those other automotive verticals. And it's easy to kind of confuse the 2. So Ouster Automotive, high-volume consumer applications in automotive. And -- so everything that we're doing is targeted at getting to volume, hundreds of thousands and millions of vehicles through series production wins. And so there are really 2 approaches. One is through strategic development agreements with automakers, where we work with the automaker, we develop a product that meets their need, get to know them, build trust and then are in pole position for winning a series production win through a standard RFI, RFP, RFQ process, bake off essentially. And the other approach to getting to that series production win is to just be in the process with an RFI, RFP, RFQ, ultimately win it and get awarded that series production win. So on the one hand, we have -- we actually have won a strategic development agreement which is a major milestone to get us kind of in the preferred position for a major series production win that will follow. And then we're also engaged with 5 series production opportunities in various stages of RFI, RFP, RFQ. And actually, notably, it's pretty hard to get to an RFQ stage, because it's a multiyear kind of conversation in a lot of cases to actually get down selected to be 1 of maybe 2 or 3 entities that are being considered for that -- well to be one awarded from the RFQ. And so yes, they've been talking with these 5 entities for over a year now in some cases. And so we've been able to accelerate the conversation by just acquiring Sense and those conversations and incorporating them in Ouster Automotive. Collectively, those 5 programs, they all hit 2025 to 2026. In aggregate, it's about 1.5 million vehicles worth of demand that we're negotiating right now, and it's over $1 billion of revenue opportunity, encompassed by them all. They all have to be determined in the next 24 months or else it's too late to hit 2025 to 2026. And we're expecting to win one or more of them actually in the next 12 months. So I feel really good about our positioning there.

And can we talk about the opportunity in terms of -- there's a lot of talk about who's got the best specs. And in most cases, they're talking about the one front facing and how far can it see and how accurate is it? I mean you've kind of talked also about maybe a sensor suite, if it gets low enough, you can put them all around the car. So when you talk about these types of deals, just solely for just one front-facing or kind of what's the potential for that kind of suite that you've talked about in the past?

Yes. And so that goes to kind of the strategy of how are we going to win these opportunities. And it comes down to 3 things: affordability, so you have to have the best pricing, because auto is incredibly price conscious and digital LiDAR is best-in-class affordability and costs; you have to have best-in-class performance or you have to at least -- you have to meet a very high performance threshold to be playing in this space; and then firmly believe you have to have multi-sensor suites. And that means not just a single forward-looking LiDAR, but actually being able to provide 3 to 5 LiDAR sensors on the vehicle and doing it for that aggressive price point. And so Ouster is uniquely positioned to provide a multisensor suite, and that goes back to this product family approach where one silicon chip powers tens of different flavors of our digital Flash LiDAR or OS series on the other hand. And then we're able to put some numbers by it. We're looking at putting 5 LiDAR sensors on a vehicle for about $1,000, whereas our competitors are really targeting a single forward-looking application, single forward-looking LiDAR for that same $1,000. And in all 5 cases that we're negotiating right now, they are for multisensor suites. Just to show that the market demand from automakers is for these multisensor suites because it enables a richer set of features in ADAS systems. It's better to have a system that can see to your left and to your right, so you can merge automatically, change lanes automatically, move through a 4-way stop automatically as opposed to just kind of automated cruise control on steroids, maybe you just have a single forward-looking LiDAR.

That's a good leading, because I was kind of curious on there's lots of flavors of ADAS and then full autonomous functionality. And clearly, when you get to Level 4, Level 5, I think LiDAR is a criteria. When you're doing 2 plus, there's still some people looking at it. solving with RADAR and image. Kind of your perspective as you talk to these OEMs, has that opinion changed in terms of including LiDAR and even something like an L2+ or some more basic ADAS?

Yes. I mean I'd love to include LiDAR and L2+. You can get incrementally better performance in an L2+ system with a forward-looking LiDAR or some other LiDARs on the vehicle, depending on where you -- what feature you're trying to augment. And we're absolutely engaged to provide that. If that's what the market wants, we're right there, and we can provide the same lighter LiDAR systems or the same performance at roughly 1/10 the cost, 1/5 to the 1/10 a cost is how we positioned it. So L2+ could be a great place to be. I think I'm personally more excited about L3, and that's what we are currently negotiating in the majority of these cases, because that's a fundamentally new feature set that has never existed in vehicles before. I think it's really, really liberating if you're able to read a book on your way to work and not actually have to pay attention. And there, you need these multisensor suites, $1,000 package for 3 to 5 LiDARs around the vehicle in order for the system to be reliable enough. But both are great, L2+, L3. Digital LiDAR is in a great position to win both.

And maybe we can talk on the software side in terms of developing a full perception suite. You were working on that pre-acquisition. I'm assuming they had some effort as well. So maybe just kind of give us a feel as to where those bolt efforts were and your ability to kind of have a software stream as you get in production?

Yes. And this has been one of our use of 3 use of funds as we went public in Q1 was to expand our software offerings. But it's a little more complicated than just building software for automotive, because our business spans industrial, smart infrastructure, robotics and automotive. We have hundreds of customers across these spaces, and there's a lot of demand for value-add software on top of our hardware, no matter who you talk to. And so absolutely, we intend to be able to provide perception offerings in automotive. My honest opinion is that we will be -- the value or the perceived value of what we're providing in automotive will be less than in some of our other verticals. And it's not a reason not to invest there, because maybe it's the difference between winning an RFQ and not to provide some additional software. But I don't think we're going to be seeing the really compelling margins from that software or revenue from that software as people might think. The days of Mobileye pulling like 80% margins back in the early 2010s, I think, is mostly over. But that same demand exists outside of automotive and industrial, smart infrastructure, robotics. And so we have kind of a diverse set of software offerings that we're working on internally. And I'm definitely excited for it. I mean there's immense opportunity to provide value-add software in this space.

And one of the challenges just getting the real LiDAR world data to kind of develop this software, obviously, chicken the egg, if you can get in some good volume SKUs, it's going to put you ahead. You saw that with the Mobileye's and the NVIDIAs. Maybe just talk about your strategy there to drive this software kind of before you really have enough cars on the road, because a lot of the internal efforts they paid for cars is just drive around aimlessly and collect data. So how do you get from Point A to Point B on your side?

Yes, I think that -- so what you're bringing up is the idea of fleet learning being necessary to build really valuable software in this space. And I think that's really hard to get to. I mean a lot of companies have been working for a decade to try to reach a fleet learning scale where that approach actually works. I'm not expecting to become a fleet learning company in automotive, to be frank. It would be great if we could, and we'll reassess as time goes on. But there are ways to provide value in the software stack with functionally safe, certified software layers that maybe don't rely on machine learning and vast amounts of kind of network data collection. And so we're focusing on things that we actually can accomplish in automotive. And then I think there is a broader discussion around creating connected solutions that are aggregating data in some of our other verticals where there are less kind of barriers in place to actually do that. So I think some of the idea of fleet learning applies across our verticals. It's obviously valuable for training machine learning and verifying it. But in automotive, I think we need to look at ways where we can provide value without having millions of vehicles on the road under our direct control.

And maybe you can talk about the manufacturing curve, you've started to -- you have the OS product lines that you're shipping. You were working on bringing in some external contract manufacturers. I don't know what Sense's kind of pipeline looks like for that kind of -- but let's just see you get one of these 5 wins. What work do you need to do to be able to manufacture in that kind of volume?

Yes. One of the things that Ouster brings to the table in this acquisition is a significant kind of -- well, significant experience with outsourced manufacturing. So we have been manufacturing all of our products fully outsourced with Benchmark Electronics for over a year now. We've had a 3-year long relationship with Benchmark, and we have a full factory running in Thailand that uses all of our custom machinery that we ship over there and then it's operated by Benchmark, while we have kind of employees overseeing the whole operation. So we fully intend to produce DF Series sensors on that line as well. So right now, it's all OS series. DF Series sensors will be produced under this manufacturing model. But in automotive, we're also open to working with Tier 1. Sometimes, once you've won business with an OEM, they may demand that you work with a preferred Tier 1, and we're going to be flexible to that. I think it's necessary in order to win as much market share as quickly as possible. So that -- it's a dual-pronged approach, but we can be a one-step shop -- one-stop shop in automotive if an automaker is comfortable with that.

I wanted to ask you on the trucking side because Ouster did have wins there, like I would say, plus where you're on the mirrors and maybe someone else is front-facing. But now I guess, thinking with the Sense products that you can now have those conversations with the trucking OEMs. So maybe just give us a perspective of where you are today in terms of design pipeline and where it could be?

Yes. So one of the things we -- so the customer you're talking about plus is a major roper trucking operator. They -- at the beginning of the year, we won a production agreement to deploy 2,000 of our sensors on the side mirrors for surrounding LiDAR on to 1,000 of their trucks that are going to their end customer, Amazon. And they are looking -- very significant, that's a binding deal with specific volumes and a specific time line that's happening in the next 2 years, and it's already well underway. And we're really focused on that kind of tangible concrete business across our other verticals. I think in the same way that we're focused on tangible concrete series production wins as the ultimate goal in automotive. And absolutely, as our partners, our customers are looking at additional LiDAR sensors to deploy on robotruck, we have a very good offering if they want a long-range LiDAR in that mix now with the DF Series or if they want additional wide field of view LiDAR. So yes, we think that we can expand our business from where it is today. And I think that very little has been determined about the broader sensor suite on some of these trucks. Like if we haven't announced or if they haven't announced specific volumes, specific time line and a binding agreement, then I think it's anybody's guess who's going to win it, and I think that we have a really good shot at all those opportunities.

I do want to ask you, I mean, it's an auto conference, but it's unique that Ouster does have such an effort outside of the auto market and potentially some of these markets can -- while not big individually, be pretty big in aggregate. Maybe you can point us to some of the verticals that you're targeting and seeing traction for LiDAR.

Yes, sure. And so there -- well, one thing I'd like to point out with the auto people is it's easy to dump on robotaxis and robotrucks and this all being later than we had hoped or expected. But that same investment, the hundreds of billions of dollars that have been invested into robotaxis and autonomy technology in the last decade, has trained a generation of engineers and motivated a generation of companies to build an ecosystem that now is at a point that it can be adopted across all these other verticals, right? The technology in robotaxis can build a fully autonomous mining machine, a fully autonomous warehouse with forklifts and yard logistics vehicles and a fully autonomous port. And so all of this technology and investment is really created an unprecedented shift in -- towards further autonomy and automation. And that's true in industrial, smart infrastructure and robotics. So great examples of this is the smart infrastructure space, everywhere that there is an intelligent traffic light or crosswalk, there's an opportunity to build a better system with a digital LiDAR, make that intersection more intelligent through the use of perception, algorithms that have been really developed for the robotaxi industry, now adopt them and make the world's most intelligent traffic light that can signal correctly in all kinds of weather conditions and understand how many people are waiting at a cross walk and turn the cross/walk light on for that. And so -- and that's an industry with huge amount of tailwinds given the infrastructure investment in the industry and also the better economics that digital LiDAR sensors can bring to roadway and infrastructure. You can just save money because you need fewer LiDAR sensors in these intersections than the camera and radar and thermal solutions that have been brought to bear in the past. So in industrials, I think that industrial is just really -- it's a huge number of vehicles that are cars or trucks, but are just working in off roadway scenarios where there's a better kind of operational design domain or a more constrained operational design domain or ODD, where there are fewer people, you can constrain the technology to a particular geographic region. And so it allows the adoption of the technology more quickly. So good examples are mines where you can just guarantee there are no people walking around, and we have a number of mining customers like Sandvik and [ TAGE ] and a number of others and with operational systems in mines across the world or a warehouse. Warehouse is a domain that's -- it's a geographically constrained to a particular building. You don't have to deal with weather, but you do have to deal with not with the complexity of a dynamic environment. And so it needs advanced autonomy, but it's also a more constrained problem than solving Level 5, all weather autonomy. So we're seeing great adoption across kind of all of these use cases.

And you track a lot of these wins. You have these, call them, SCAs, where you kind of -- the number of customers that you have these strategic agreements. But at the of the day, you still going to wait for things to go to production volumes. I'm curious, it's easy with autos, right? There's a dozen auto manufacturers or maybe 20 that you can track. And probably you know what's going on and each one is fairly big. I mean, maybe just talk about the channel in these broader markets. And then kind of how do you track it and handicap what's going to actually go to production? And what's your kind of visibility as you look into next year and the year after?

Yes. And so this is definitely a challenge. We're still early. On the one hand, the investment in autonomy is immense across the entire global supply chain of the world. And no company is debating whether this is an investment that's worth making. And you see almost every day I see an article about Amazon investing in increased supply chain automation and other big companies doing the same. And so the case is closed on this being -- there being a significant demand for this and there being a business case. But there's a diversity of maturity in the end offerings. And so one of the things that we try to do is a conservative approach of papering the contracts we have with leading customers. And that's our strategic customer agreement approach, where any customer that we deem is like something -- a customer that has production intent, that's going to continue to purchase over time, we negotiate a contract with them called the SCA, and it encompasses negotiated pricing, 3- to 5-year forecast direct from the customer, terms and conditions on point-of-sale and things like that, and specific ouster products and specific kind of targeted application for those products at the customer. And we signed 62 of these to-date. And in aggregate, they're collectively worth about $470 million in contracted revenue opportunity. And this is our way of building confidence in our own pipeline and also being able to commute something that is consistent to the market for investors and analysts alike. And we've talked a lot about this plan. But we're going to continue with this. it's a pretty conservative approach because it requires us negotiating a contract with every one of these customers, and we have 62 now. But then we have to qualify a much broader pipeline of tens of thousands of potential customers, thousands of which we may be in direct -- potential customers, thousands of which we may be in conversations with at any one point.

And I'm kind of curious also just in terms of -- you can ship a product, I'm assuming tomorrow if it's an OS series, right? I think not every LiDAR company can say that. In fact, some have road maps that are very public that won't be for the next year or 2. So I'm kind of curious when you compete, how many players do you really actually see? I mean, I think people look at the landscape and say, "Man, there's dozens, tens of dozens of companies." How to really sort through it? Maybe you can give us a perspective of is it really that wide? And I guess how many can actually shift to that?

Yes. And that's -- I mean, that's really the -- Ouster is unique because of our digital technology, our diversified strategy across markets. But at the end of the day, that's only meaningful if we can execute on it. And that's really where we have stood out additionally as we started the company in 2015, same time as all these other companies that have just gone public. And yet we have the most revenue, the most customers, the highest margins, the most units shipped, the most generations of products released of any company started in that time frame. And so the ability to execute and actually deliver on this whole strategy that I've played forth is ultimately the determining factor between success and failure. And so you're right, in the market today, we don't compete with almost -- it's 5 or under companies. It's really the incumbents and a couple of companies that have copied the incumbent technology, analog LiDAR technology. And it's fewer than 5 companies commonly. So there's a lot of noise in the market about an announcement with a partner, partner announcements and things like that. But the very few companies are going beyond that and actually announcing production level agreements with customers like Ouster is doing with our SCAs.

With that, time flies, we're out of time. I really appreciate it. Talk to you soon, Angus.

Thanks so much, Blayne.