Innoviz Technologies Ltd. (INVZ) Earnings Call Transcript & Summary

Good morning, and welcome to Innoviz Fourth Quarter 2021 Earnings Conference Call. Joining us today are Omer Keilaf, Chief Executive Officer; and Eldar Cegla, Chief Financial Officer. [Operator Instructions] I would like to remind everyone that this call is being recorded and will be available on the Investor Relations section of our website at ir.innoviz.tech. Before we begin, we would like to remind you that our discussions today will include forward-looking statements that are subject to risks and uncertainties relating to future events and the future financial performance of Innoviz. Actual results could differ materially from those anticipated in the forward-looking statements. Forward-looking statements made today speak only to our expectations as of today, and we undertake no obligation to publicly update or revise them. For a discussion of some of the important risk factors that could cause actual results to differ materially from any forward-looking statements, please see the Risk Factors section of our Form 20-F filed with the SEC on April 21, 2021. I will now turn the call over to Omer. Please go ahead.

Hi, everyone, and thank you very much for joining us. I'm happy to host our fourth quarter earnings call. And with us today I have Eldar Cegla, as usually, joining me, our CFO. And today we have a special guest, which is Oren Buskila, who is one of the founders and is the Chief R&D. And he will talk about our R&D progress to date. With that, I will start the presentation. Okay. I want to start with talking about how Innoviz have met our 2021 targets. We had a design win with Level 4 with an automotive company for a shuttle program using multiples of our LiDARs, a shuttle program that is still due to launch end of this year or beginning of the following months. This program is making good progress, and we are very happy to be part of it. This program is using InnovizOne. Last March we announced a new product innovis2 InnovizTwo with a very significant performance improvement and cost reduction, and we told the market that we're going to have first samples by the end of the year, and we have managed to meet that target. We were qualified as a direct supplier for one of the largest car makers in the market, as part of us competing on the program, which we are still expecting decisions to be made soon. And we are going -- we are following the same process as other car makers today. We ramped up a first automotive-grade, high-volume production of InnovizOne with our partner Magna to support the BMW program and shuttle program for automotive grade [ body ]. Additionally, on that, we talked about having to reach a design freeze of InnovizOne and I'm happy to say that we have managed to meet that target. Now, moving our efforts to the production validation and prepare to the launch of the programs. Additional to these efforts, we've included new capabilities in the company. And opposed to the InnovizOne effort, we did the design manufacturing -- design for volume manufacturing with partner with InnovizTwo and now Innoviz360. We do the design for manufacturing ourselves, and we have our own NPI introduction line for small volume production of the InnovizTwo in our own facilities, which gives us the ability to move very, very fast. That's what allowed us to have first samples so fast. We've announced several announcements on new partners and customers through the entire year of '21, and we will continue to do so. Last earnings call, we talked about the many efforts that we're doing to meet our targets to have InnovizTwo available and show it for the first time in the Consumer Electronics Show in Vegas. And I'm happy to show you now a video of the InnovizTwo first samples. [Presentation]

We have a working product. With that I want to talk about what we managed to accomplish in the last quarter. We announced 5 new contracts for pre-production deals, and I will elaborate on that in the following slide. We continue to work on several RFIs and RFQs in the market, and we have very good momentum. One of the programs that we were hoping to announce end of last year has been pushed by the customer for all these days. They've asked to make some changes in the requirements. We have responded to those requirements, and we are now waiting for their decision. We are very confident in our ability to secure very meaningful business this year. There are additional RFQs coming along and with our portfolio, I believe that we will be able to announce more this year. We increased our order book by USD 200 million, following progress we have in the shuttle program decision in order to include more LiDARs per vehicle and the maturity of the program allows us to increase our expected revenues in this program. We have our first samples of InnovizTwo that allow our customer to test the product and make a decision. We announced a new product, a very innovative one, Innoviz360. It's a new category which allows Innoviz to participate and compete on many additional programs, and I will elaborate on it later. We had a webinar for Level 3 standardization where I introduced the white paper we created to talk about what are the requirements for LiDAR to meet Level 3 performance, and we see Innoviz as a strong player that can actually help in consolidating the requirements for this market. We had a session in MIT course for deep learning showing the progress of our perception software. We were awarded with the IATF certification for production quality for automotive. Other than that, we announced our new partners and customers just recently. On the 5 contracted pre-production deals, we announced a new customer, Obayashi, one of the largest construction companies in the world using our LiDARs for automated driving. Another customer is Syracuse who is also using our LiDARs for automated crane control. You can see here a video they created using InnovizOne. It's an Israel company we're very happy to work with. Whale Dynamics is a company we announced a quarter ago about our collaboration and now reaching an agreement in us supporting their efforts to bring platforms - autonomous Level 4 platforms to the market. We also agreed to -- we reached an agreement with an Asian company to bring automation [indiscernible] with InnovizOne and Twinner to support their efforts in automation in defect detection, again using InnovizOne. We are working to increase our customer work through the maturity of InnovizOne and soon also our with InnovizTwo. Our funnel, so we have -- until today, we have 2 series production programs with 2 big automotive companies, one is BMW. We host Richard Rau last quarter talking about the focus of the program and more details will be shared throughout the year about the launch of the program. And the shuttle program which I referred earlier. We increased our pre-production program by 5 deals as I described earlier. We have currently 12 companies in different stages of RFIs and RFQs, and many of them reflect very meaningful opportunities of order books. Our targets for '22 is to reach 10 pre-production programs that will allow us to meet our revenue targets and at least one more automotive program that will allow us to increase our order book by at least 30%. Now I would like to talk with you about Innoviz strategy related to our business development and automotive and beyond. The way Innoviz is operating is that at every point of time we look at the total addressable market of LiDARs. There are different markets that LiDAR could be used, but it's always very important to identify which of the markets are more likely to go through an [ avalanche ] sooner. We, by doing that study, the requirements, understand what are the frictions, what are the requirements, the needs, and wield the right chipset you would say or the building blocks that allow us to be -- to provide the best fit. We industrialize those platforms, and we continue to improve them and also provide the right software tools. To give an example to that, today there are different markets for LiDARs, but the fastest-growing market for LiDARs today is L2 and L3. You see the different car companies that are doing -- they are looking for LiDARs for launching Level 2, Level 3, and we see that as the most interesting LiDAR opportunity in the coming years. It's also interesting to understand that the decision-making of those programs are these days. In the next year or so, most of the decisions for this market will be taken and the car companies that will make decisions following that would rely on the decisions that these carmakers have done. We've studied L2 and L3 very thoroughly. We just introduced a white paper that shows how deeply -- how deep is our understanding in this field, and you can actually find that on our website. And for that we introduced Innoviz -- InnovizTwo meets all of the requirements we got from our customer alongside with our perception software which comes with a lot of maturity, a lot of experience coming from InnovizOne. InnovizOne has hundreds of thousands of kilometers already driven to try to find any issue that might come along, the ability to show that experience reduces the risk for any car maker that wants to make a decision today on a technology that can allow them to launch safely would go for InnovizTwo in [indiscernible]. The next market that we identify is an interesting one is the Level 4. There are actually different RFQs that we are starting to come along with the different customers, different prospects of this market, could be shuttles or robotaxis or trucks. And that is expected to launch in '26 or '27, and we are working with those customers. The requirements for Level 4 is slightly different than Level 2 and Level 3. There are a lot of attention for industrial design because it's not only one LiDAR for looking. There are multiple LiDARs around the vehicle, and it's very important to reach a zero blind spot configuration. For that application, we introduced Innoviz360. We designed it over the course of the last year, and we introduced it just a couple of months ago. And we see that as the right combination of InnovizTwo and Innoviz360 to serve those markets. Using our perception software on top of it provide our customers to grow from Level 2 to Level 3 and now to Level 4. Beyond automotive and, of course, automotive aggregates very different applications. Also, military and surveillance and construction sites and those aggregated opportunities will eventually surpass the automotive market. Every application today that is using today 2D sensors will eventually would transition to a 3D sensor. And as such, we are learning the requirements of this market which is much more fragmented, which requires much better seamless integration. It's no longer an automotive process of integration for several years and requires several analytics tools that we are building. We are looking in those markets and we are learning what additional platforms and LiDARs are required, and we will continue to provide new products to sell those very interesting opportunities, and we will have them ready for the decision making on when this s-curve of revenue generation will arrive. In terms of software development, and we've been developing our perception software for about 6 years now from the inception of Innoviz, we identified the need for perception software from the first day. It was very clear to us that any customer that is requiring a LiDAR would not be able to use just a sensor as a raw data but will require the object detection classification, the movement detection, guardrails, landmarking, et cetera. And that's what we offer with the LiDAR and price it alongside the LiDAR. That will generate us revenues for software, not only as a onetime fee but also as a subscription-based support for after the launch. So for Level 4 and Level 5 and those are RFQs that we are looking at. There are additional requirements that we need to build which are related to multi-driving direction and sensor fusion and, of course, supporting the zero blind spot, and that will also come on top of the pricing of the LiDAR. For the non-automotive, the perception software that we already developed provides us the best, I would say, incremental step into this market, having a mature automotive-grade object detection and classification gives us an ability to penetrate this market very fast. We will build additional tools, and the fee would be defined per application eventually. Our cash position is very strong. We have USD 370 million just raised in 2021, with USD 300 million in our bank accounts starting this year. We are very conservative in the way that we manage our funding. As an example to that, we already foresee all of the planning that we need to have for our build this year and sourced all of the components so we'll not be sensitive for the chip shortage in this year. To summarize, Innoviz ability to succeed in this market and to lead it, we have 2 significant old customers that give us a very large order book and a very large pipeline ahead of us that we see a very good chances to take. We have a cutting-edge technology with the proven capabilities and a mature one that allows carmakers to have zero risk in taking that technology. We have the perception software probably with more mileage than anyone in this category, with the capital to execute our strategy, we have the industrialization experience, and we have today 3 production lines to support it, and we have the automotive experience that is baked into everything we do. Of course, we have also the production -- our product line which will allow all of our customers to grow with us from Level 2 to Level 3 to Level 4 and 5. And with that, we are going to grow the company and continue our efforts. I will hand over to Oren Buskila to present our R&D efforts.

Thank you, Omer. I'm very glad to be here and really excited to share this update on the R&D progress in the past quarter. And I'll start with taking a step back and looking at the 6 years since Innoviz started. Now we started back in 2016 where the only LiDARs available for the automotive market were of very poor performance. Very far from reaching the requirements of the automakers for self-driving, and that was back in 2016. And since then we've made an exceptional progress with the InnovizOne. In 2020, we launched the first [ super still ] of InnovizOne which provided unprecedented levels of performance both in terms of range and in terms of resolution, making the LiDAR finally suitable for automotive self-driving on public roads. And during that time, we did not neglect any of the requirements dictated by the auto industry for reliability, quality, cost, et cetera. And now in 2022, just 6 years from the inception of the company, we're ready to take it to the next level with the InnovizTwo. The InnovizTwo will be a much-improved version or much-improved generation after the InnovizOne which will have a dramatic boosted performance, but at the same time will have a dramatic cost reduction. In terms of performance, the InnovizTwo will have twice the resolution in each of the axis, so 0.05 by 0.05 degree in X and Y axis. It will have a wider field of view and taller field of view of 120x40 degrees and all at the same time it will have a longer range which is only more difficult once the resolution is finer, and this will come at the dimensions and the price point fit for Level 2 consumer cars through Level 5. Now for some more details on what we've been doing in the project of the InnovizTwo in the last quarter. So as Omer mentioned, we actually met the objective of delivering the first InnovizTwo samples back in January. That was at CES Vegas where we presented the B0 samples, the preliminary samples of the InnovizTwo. Preliminary as they are, they met most of the performance objectives targeted for the InnovizTwo already in the first sample, including the field of view of 120x40 degrees, the resolution of 0.05x0.05, and the range of more than 200 meters for a very dark 10% reflectivity target. That was achieved with a frame rate of 10 frames per second, which is going to improve in the next versions of the product. Speaking of the next versions, the next one is the B1. The InnovizTwo B1 will be available in Q3 of this year, and it will feature a massive decrease in the size. So here I'm holding in my hands, and I hope you can see them. This is the InnovizTwo B0 that's now sampling. This is the Innoviz B1 coming up later this year. And you can see both from the presentation and maybe from the video that we've reduced the height of the B1 dramatically versus the B0, with the target of making it suitable for more diverse mounting positions on the car, including the roofline where the vertical height is a critical dimension. The field of view was adjusted...

Show them the [ video ] of the car.

Thank you, Omer, for reminding me. This is try and find the LiDAR here. This is right here in the roofline. This is a suggested concept design from Innoviz of how to integrate the LiDAR with the car design, and you can see a very seamless, very sleek integration with the car's exterior. The field of view of the B1 was adjusted in order to accommodate the thinner height dimension. The frame rate, we've been able to meet the production-intense frame rate with the B1 which is 20 frames per second, while not sacrificing neither the resolution or the range of the B0. The B1 also is designed for manufacturing, meaning that all of its parts, components are already designed to be made in volume with high yield and low cost. And in parallel to developing the B1, we're also developing and bringing up the production tools required for the production line of the B1. Of course, this is a low-volume production line later to be replaced by the mass production line of the future versions. On the top right you can see a point cloud taken with the B0 just I think yesterday, and this is really recent. And those of you who are used to looking at point clouds I think will probably admire the resolution and the angular accuracy. This is really a spectacular performance that we're already seeing from the B0 and this is only going to improve with the B1. The InnovizTwo will also be the first product from Innoviz to include onboard perception software. And this means that we are now completing the coverage of both common perception architecture of cars, the first one being distributed processing architecture and the second one is centralized processing architecture. In the distributed architecture, every sensor processes its own raw data locally and outputs just an object list of what it detects. So the InnovizTwo will run this software, this computer vision software on the LiDAR box itself inside it, thus eliminating the need for any exterior processing of the point cloud. We will, nevertheless, output both the object list and the raw point cloud in case this is of use to a customer. Now this local processing reduces the compute load on the central computer of the car, simplifies the data interface, and reduces the latency, all making the integration of the InnovizTwo into the car system much simpler and also reduces overall costs of parts in the car. The second architecture which we've all supported also in the past is the centralized architecture where all the sensors in the car, including the LiDAR, do not process their data locally but rather transported into the central computer which does the processing and the fusion of the perception features. Now Innoviz does not just output the point cloud even in this case but also provides the perception software to run on the central compute platform, meaning on a different chip than what resides inside the InnovizTwo. The output is a raw point cloud and the Innoviz software running on the central compute is responsible for detecting objects on the road as you can see in the image to the right. Now we support most of the processors from various automotive-grade vendors today, and we intend to increase the -- to expand the support to further vendors. Now I'd like to show a quick video of the InnovizTwo B0 here in beautiful Tel Aviv. This was also taken not very not late ago.

It's not the most updated.

It's not even the most up-to-date point cloud. We are making progress every day recently, but what you can already see here is the exceptional resolution and range that is visible here. And this is very, very important to note that this does not include any aggregation of frames. Each frame here that you see was taken separately with live recording from the LiDAR. So that was how Innoviz is disrupting the automotive LiDAR space. But the other segments, other than automotive, well, up to today underserved by relatively low-performance LiDARs, which are also high-cost and typically quite big. And this is soon going to be changed with the Innoviz360, the first LiDAR from Innoviz which will have a 360 degrees coverage, a surround-view coverage of the LiDAR. The vertical field of view will also be very high, 64 degrees. And all this will come in a very small package. You can see actually a mock-up, at this point, right here, which is very small compared with other LiDARs available today. And the cost will also be exceptional compared with any other comparison today. Now the way that we're able to offer both this performance, this cost, and the very aggressive timeline for samples is by leveraging the same components, the same optical engine that we have in the InnovizTwo, we're going to integrate that into the Innoviz360. We're designing the Innoviz360 around it while only adjusting the scanning mechanism for 360 coverage. Now it's really important to understand why the resolution is so important, why it matters that the Innoviz360 will have that resolution which is higher than the full high definition. So what you see on the right and the left are actually the same scene but captured in simulation by 2 different LiDARs. On the left is probably the best-in-class LiDAR available today, a spinner LiDAR. And on the right, you can see the future point cloud of the Innoviz360 which is a factor of 10% in the...

10x.

Thank you, Omer. 10x the number of lines. And why is that important? So when looking just at the center of the field of view, well, where the lines are dense even in the competitor LiDAR, then most objects are quite visible in the center of the field of view. But as we look towards the edges of the field of view, you can see that with few scan lines there is large -- very large spaces between the lines making it possible to miss small objects which might be critical for the car to detect. So on the right you can see a dog and a cat, stimulated ones, of course, are easily detectable by the coverage of the InnovizTwo, whereas on the left you can see that they're very hard to detect, actually quite not seen. Innoviz is working not only on the system level but also on the underlying components such as the ones that now are driving the InnovizOne. We are now working on the next generation of those components, which will drive the future products. The first one being the next-generation ASIC, which will support an increased pixel rate, an increased maximum range of up to 500 meters, and will also have support for additional interfaces and future security standards, which are going to be adopted in the auto industry. Samples of this ASIC are going to be available in 2023 and will support both the future products and the existing products. This is all backwards compatible with the current LiDARs from Innoviz. The second component that we're working on is the next-generation detector which will have the best-in-class sensitivity of all detectors of its kind, enabling increased range, improved resolution, and some other benefits which we will not expand on now. Engineering samples are also going to be available in 2023. Next update is on the perception software. We are now at the stage -- we're now at the stage of it's called SRS or safety-related scenarios. These are staged scenes, emulating actual roads, things that cars -- our customers' cars might encounter, and the purpose is to validate both the point cloud performance and the performance of our perception features. The other major effort we're now working on in the area of perception is the endurance runs. This is a massive, massive worldwide campaign for collecting hundreds of thousands of driven miles in several countries towards the validation of again the point cloud and the perception features in real-world conditions. This only began once we've met the KPIs, the performance necessary for production. Now we're ready with our customer to start racking up the miles of this recording campaign. To summarize, the Innoviz R&D objectives for this year are first to complete the product validation of the InnovizOne towards the beginning of the volume production of it; second is to deliver the B1 sample of the InnovizTwo which is designed for manufacturing; third is to reach the first engineering samples of Innoviz360; and lastly, to maintain and increase Innoviz technology leadership with our components and perception software. And with that, I wish to thank you and pass on the presentation to Eldar Cegla, our CFO, who will present the 2021 financial results. Eldar?

Thank you very much, Oren, and good morning, everyone. We'll start in a moment. One moment, please. We have it, yes. So Innoviz 2021 -- in 2021, Innoviz successfully closed a business combination with a SPAC which provided over USD 370 million in gross proceeds. At year-end, we maintained a high liquidity level of USD 304 million in cash, short-term deposits, and marketable securities. Our operating cash flow during 2021 was within the planned budget and Innoviz's debt-free cash asset enable it to execute its 2022 plans. Revenues for 2021 were USD 5.5 million, while 2020 revenues were minus USD 9.4 million. InnovizOne-related revenues in 2021 increased by 28% to USD 4.3 million compared to USD 3.3 million of total annual revenues in 2020. The company expects InnovizOne sales to continue to increase this year, and we are also targeting to sell the first samples of InnovizTwo. Operating expenses for 2021 were USD 152 million, an increase from USD 66.2 million in 2020. Operating expenses in 2021 included USD 64.7 million of stock-based compensation compared to USD 3.2 million of stock-based compensation in 2020. The increase in operating expenses compared to 2020 was primarily due to the increase in stock-based compensation and people-related expenses. Additionally, research and development expenses for 2021 were USD 93.3 million, an increase from USD 57 million in 2020. Research and development expenses in 2021 included USD 25.5 million dollars attributable to stock-based compensation compared to USD 2.6 million attributable to stock-based compensation in 2020. To conclude, we will continue to focus on developing groundbreaking technology and products for the automotive market and beyond. We believe that we are well positioned to execute our strategy and growth plans this year. And with that, I will turn the call back over to Omer for closing remarks. Thank you.

Yes. Thank you, Eldar. So again, thank you very much for joining us this day. Innoviz is working very strongly -- working very hard to succeed. We have I think the best setup that we can wish for with very promising products and portfolio. And our team is working very hard to make that happen. With that, I would like to invite you to ask questions. I'm happy to answer them.

[Operator Instructions] Our first question today is coming from the line of Mark Delaney from Goldman Sachs.

Yes. I was hoping, first, if you could talk a little bit more around the top line trajectory for 2022. I realize you're not giving specific guidance, but maybe you can talk about whether or not you think there's potential to begin series production this year. You have the BMW program that I thought you had the potential to start in '22. So you guys were thinking about revenue in 2022. Is that something we should be contemplating in the latter part of this year?

Yes. I'll start and maybe Eldar can continue. So the launch of the programs due to the announcement of our customers, which we obviously cannot talk about, we expect that -- we are under the assumption that the launch of the shuttle program is done this year and BMW I leave it to them to announce. And I believe that in the near future you'll get more information about that. There are several customers that we are serving beyond supporting those programs. And some of those that we described today will allow us to meet a big portion of our target, and we are going to continue and bring those deals to the table. The fact that we have InnovizOne now more mature and being able to design the -- freeze the design and bring up the production volume helps us to give more customers in their hands the product to make a final decision and move to an order, which will allow us to meet our targets. I think, generally speaking, the programs that we've been competing on and where we're actually expecting them to be made decision, end of last year were pushed to this I think about 6 months of delay that was in the market is pushing the revenues by about 6 months, but the ramp-up and the growth of the revenues are expected to be in the same manner that we initially assumed.

Yes. So adding to that, A, our focus is on market share -- winning the market share. As Omer said, there is a shift in revenues from last year to this year once we win those -- hopefully, be able to win those programs and they will generate additional revenues. We also plan to have the first samples of InnovizTwo starting to sell and roll out into the market.

I want to add that the fact that we included now Innoviz360 to our portfolio is opening a new TAM for us and new opportunities. We expect to be able to penetrate the non-automotive market in the following year and grow our potential.

Okay. Just to clarify, I realize we need to hear from BMW on when they would start series production, but would Innoviz be ready to support them this year if they do choose to begin later in '22?

Yes.

Okay. That's helpful. My follow-up question is on ADAS L2, L3 program with a big OEM customer you talked about it on the last call and you said today they've asked for some changes and you guys are working to address that feedback and still feel good about winning it. Maybe you could elaborate a little bit more on that in terms of timelines and how long do you think this may take. And you mentioned a new variant of the InnovizTwo, the B1 that will be ready in the third quarter. Is that the variant of InnovizTwo that you guys would need in order to potentially win that? And so we're really thinking it's not until 3Q that perhaps we'll get more of an update on that front.

I'll start actually with the second question because I think it's actually you assume correctly. We designed InnovizTwo following discussions with our customers and the customer that we expect to make a decision soon has a very, I would say, short timeline to make a decision and launch the program. And we are trying to make sure that we are going to follow all of their requirements throughout the year and supporting them with the product they are looking for, and that's what -- the changes that we've made also to the B1 sample versus the B0. And to your first question, yes, we have very good confidence in that program. It's over a 1.5 year I think, even more than that today, very close discussions beyond the negotiation of the financials and the technical. And we feel still very good position to possibly win it. In regards to when it's going to happen, basically, we are waiting for a decision. When it's going to be made, it's not in my -- I cannot tell. But we are hoping to hear a decision soon.

Okay. I guess if the sample of the B1 is in the third quarter, if I heard that correctly, could you get a decision from that customer before they sample it in 3Q, or do they need to have that in hand in the third quarter before they make that decision?

No, they don't need -- the B0 is already meeting their needs in order to make a decision.

Our next question comes from the line of Michael Filatov from Berenberg.

So first one just branching off of Mark's questions. In relation to that large OEM that you're engaged with right now, I think you previously said you were competing against one other supplier for that business. Has that changed at all? Is there a change in the competitive dynamics with that particular engagement?

Well, the requirements have changed since the end of last year. And as far as we understand, the other competitor was not able to fulfill them. I'm not aware of the situation right now whether there is another competitor out there that is able to fulfill it. Our assumption that it's not, but we don't have the full -- probably. We don't have the full visibility of the landscape.

Okay, understood. And looking at your guidance to potential order book growth of 30% or more. I guess if you assume that you did win another OEM program, whether it's this one we were just discussing or another one. I guess I would have thought that a win with another large OEM would impact the larger -- the order book to a larger degree than maybe 30%, just given the size of the BMW order book. So maybe you could just talk a little bit about that? Is there an assumption of an OEM win in that order book growth of 30%, or is that without an OEM?

Well, an OEM win would make that much higher. We did like an weighted average of the different opportunities. But, yes, an automotive-grade program win would be higher than that.

Okay, great. And then, again, I know you haven't talked about BMW's launch timing, but let's just say hypothetically if you were on an OEM program that was going to launch in maybe early 2023, would you start to see revenues from that program in late 2022? I guess would there be some pre-production revenues associated with that program or ramp up in revenues associated with that program in the quarters preceding launch?

We are already making revenues from the BMW program, as you say. The pre-production testing requires validation of hundreds of units will generate revenues for Innoviz. So the answer is yes.

Okay, great. And just one more. In terms of we talked about this one OEM program that you're engaged with now, but it sounds like there are several others. How many are you engaged with right now in maybe RFQ process or late-stage negotiations, and how many do you think will be decided this year if you had to guess?

There are currently about 6 OEMs in discussions out of the 12 that we mentioned. And I would say that this year we expect to have at least half of them making a decision or in the second half next year, but at least 3 other than the 1 that we're talking about.

Our next question comes from the line of Andres Sheppard from Cantor.

Congrats on the quarter. A lot of the questions have been answered. Maybe one quick follow-up is you've mentioned the 10 pre-production programs you're expecting to secure in 2022. I'm wondering can you quantify those a little bit further in terms of what those could mean for both top-line revenue and maybe how many of those are in late stages, middle stages, early stages.

So there are several customers that we're already in discussion, as you said -- as I mentioned earlier. There are 12 customers, 6 of them are in the automotive, 6 of them are not in the automotive. And I would say that in most of those -- to reach those 10 this year, we are already in, I would say, late discussion stages. And we see those are as important in order to fulfill our targets for '23 in terms of generating revenues. At least a big portion of what we expect as revenue generation for '23.

And maybe one last follow-up for me is on your liquidity status, right. So if we just look at cash, cash alone, I think that number decreased rather substantially for the quarter. And so do you anticipate to having to go into the market to raise additional funds? Does that -- does this new cash position, does that change the previous strategy? Any color around that?

So the cash level decreased by roughly USD 20 million or USD 21 million which is within our plans. So it's not more than what we have thought. Additional capital raise has to come with the context. It's not a standalone. If there is meaningful business, if there is meaningful transaction that creates value for us, for our shareholders, for the business that we are bringing, that might be an option. But currently, at this current stage, we have raised significant amounts of cash last year, and we are fully prepared to execute our strategy this year.

All right. Maybe a quick follow-up there if I can. So that makes sense. But if we just look at cash and equivalents, I think we went from USD 140 million in Q3 to about USD 25 million in Q4. And so what is a comfortable level of just cash and equivalents that you feel comfortable going forward, and is there a threshold that you wouldn't want to get below? Obviously, you don't want to have negative, but are you comfortable with a USD 25 million in cash going forward?

I'm not sure I follow the question. The total amount of cash that there is in the short-term deposit and marketable security is USD 304 million.

Without debt.

Without any debt. And last quarter in Q4 we spent around USD 20 million, a little more than USD 20 million. So in terms of cash levels, we feel very comfortable. So I'm not I hope I'm answering your question.

That's helpful. Yes, that answers the question.

Our next question comes from the line of Mark Delaney from Goldman Sachs.

Thank you for taking the follow-up. I was hoping to dig in a little bit more on that last topic of cash flow. Do you have any expectations for cash use in 2022 that you're able to share with us?

I can maybe talk about this. Yes, so I would say that our plans for 2022 is split into, I would say, 4 main activities, and they are split and I would say about 1/3 comes from effort to bring InnovizOne to complete the production validation and testing and ramping it up, readiness for the launch of the programs, expected at the end of the year; another 1/3 related to activities in bringing up InnovizTwo to volume production for 2023. We want to be prepared to start ramp up volume and have additional volume sales in '23 coming also from InnovizTwo. And then I would say out of the last 1/3 I would say that 80% of it comes from expenses related to the perception software and software development and the rest is related to activities in the bring up of Innoviz360 and additional, I would say, innovation in the CTO office, in the innovation team, roughly speaking.

Got it. Yes, and any sort of overall quantification around cash use in 2022? What does that add up to cash use, USD 100 million, USD 150 million. I'm just trying to get a sense for the cash usage this year if you're able to give some rough ranges on that.

So we are not talking about the forward-looking cash use but 70% of our expenditure is the R&D. So you saw the amount of cash that we used last quarter, there will be some increase obviously. And again, as I said, we are very well positioned in order to be able to execute our plans for this year and have the R&D fully funded and execute the main targets that Omer mentioned.

Maybe I can add to this, Mark. I think the reason that we are hesitant in really giving right now just numbers because there are several programs that we're competing on and those are very big ones. And based on the activity that we will need to bring up the different teams to support the different programs would require us to act differently, and that's why we're going to be very efficient in the way that we manage our expenses, but we are really looking for a very exciting year. So with that, we will just -- we will manage it to bring our best results.

Okay. No, that makes sense.

Our next question comes from the line of Michael Filatov from Berenberg.

I just have another follow-up. Since you've got Oren on the line, I figured I'd ask him a question. So you talked about the update to the ASIC chip and to the receiver that would be expected in 2023. Is that -- would those updates be built into whatever products that you're now offering or potentially offering to customers such as this large OEM you're engaged with? Would those be retrofitted into the design, or would that be on a next-generation product that you would be competing with?

Okay. So these 2 components mostly targeted towards the next-generation products, meaning that the upcoming products of InnovizTwo and the Innoviz360 are not dependent on those components in order to meet their performance requirements and schedule. However, they will be able to benefit from them in the future. So we may have versions of the InnovizTwo and the Innoviz360 with even enhanced performance. But again we're not relying on them in the short term. The InnovizTwo and Innoviz360 can reach the market even without those projects.

It is always related to the design freeze that you have for certain platform. Once you reach a certain point, you can't change anything in the hardware. We've met that point with the BMW program somewhere about a year ago. And basically, once you reach that point, you only need to make sure that you finish all of the design validation and make sure nothing needs to be changed, but you're not allowed to change any meaningful components. So it really depends on the timing of the program, and by that, we will be able to decide whether this shift is relevant or not. Specifically for the programs that we're offering today for launch in '24, we're not assuming those are required. It is really dependent on the launch of the program and the timing of the program and the flexibility of the customer. For the non-automotive market, those we'll be able to leverage and those advantages obviously much quicker. And I hope that answers the question.

Yes, understood. And just one more on the software, since you touched on it today. Are you guys are you just doing the perception software that basically processes the raw LiDAR point cloud data or are you offering some sort of holistic perception stack that maybe tier 1s or OEMs will use? And you talked about licensing that software. Is that included in say USD 500 ASP that you referred to for the InnovizTwo or sub-USD 500 ASP, or is that on top of that? And what would that software licensing revenue opportunity be?

Sure. I'll start with the second question. It's easy. So that comes on top of the pricing of the LiDAR. And, of course, if there are 2 LiDARs in a vehicle, then we only price it once. It depends on the application. And then you have the subscription fee for support, whether there are requirements for supporting, adding additional features post the production. So that's also something to include. In regards to the capabilities of the software and what we offer, so Innoviz is not developing driving decisions. Those are done by our customers. We are offering the perception software, which allows our customers to have a seamless integration of object detection and classification. Basically, what you can think about in terms of what we offer, carmakers wants to have a platform that has the computer vision capabilities of the camera and the same interface, the exact, really copy exact capabilities from another sensor that provides its redundancy, just that it's a sensor that does not degrade when the camera degrades. Meaning that the interface is, I would say, almost ignorant of what's behind it. You'll get an object detection classification, movement detection, drivable area coming from our LiDAR, that's what we offer. That's the package that we provide, and we price it on top of the LiDAR. And that's what's allowing our customers to reach Level 3 or higher, and that's priced on top of the LiDAR. I hope that answers your question. And if not, please let me know.

It does. Can you quantify the software opportunity, like the dollar content or...

About USD 50.

Okay.

Yes. Our assumption is that we use this very similar business model of Mobileye at the time where -- basically what Mobileye provides the market is not a camera. They're offering computer vision. The camera is done by the tier 1, Mobileye provides the chip with the computer vision which basically translates the 2D image to object detection and classification used by customers. They don't want to use the raw 2D data. They want to use the object detection, the insights the sensor is providing. So that's -- I think the industry has converged to a certain point of willingness to pay for such a product which is object detection and classification over a camera which is about USD 50, and we have the assumption that the same willingness to pay is going to be for the LiDAR and that's how we price it. So the USD 50 comes from the object detection and classification that we provide. Only that is done on our own sensor, which is the 3D sensor.

Yes, understood. And just have you seen the take rate? You assume like is BMW using that software, what would you expect other OEMs that you're engaged with right now to all be using that software or what does it look like so far?

Almost all of our customers are relying on the ability to get our perception software from us.

Understood.

Our final question for today comes from the line of Kevin Cassidy from Rosenblatt Security.

Congratulations on the progress. Maybe just to expand on the perception software. Will you be opening that up for third-party software people to write software to your LiDAR?

Yes. Well, our customers have the option to either use our own perception software or to use either their own or third-party software. So some customers, especially in the, I'd say, robotaxi vertical, for example, do choose to use their own software or some integration with a third party. So we're not blocking anyone from using our point cloud, our raw point cloud with their own perception software or third party, actually even encouraging it. But many customers, especially in the OEM space, do prefer to get that as a package and solution together with perception [ formulas ].

Maybe I'll just add to this. In automotive, the perception software also needs to go through a very rigorous process of [indiscernible] a spice which means that obviously you cannot just let anyone to code in your processing. It has to be managed in a very good way. And therefore, in automotive I think very less likely that we will be able to host any processing code in our processing unit, but as Oren said, once the data -- we do provide the raw data for centralized computing where the code runs on the main computer of the car, then someone can take the raw data and try to extrapolate features out of it. Even if it's done in parallel to our software that runs on the centralized platform, which we are doing. So we have the perception software which is embedded in the LiDAR in InnovizTwo, but we also are porting those algorithms. For example, we announced NVidia collaboration in terms of porting all of our algorithms from the ASIC -- from the chip that we have in the LiDAR to the NVidia platform because there are platforms that are doing centralized computing and the interface is raw data, so we have to port that processing into that platform. So we're not blocking any architecture. But, of course, when it comes to automotive, it's much more conservative in that manner.

I see. And maybe just to understand the magnitude of this. What percentage of your engineers are software engineers working on this?

So I would say there's software engineers in general and then there's the pure software organization. I'd say the perception organization is roughly 1/4 to 1/3 third of the R&D headcount.

It's about almost 100.

Yes.

Almost 100 people working on the perception software. It's funny because Innoviz is very known for the LiDARs, I mean, I wish. But a very strong, I would say, growth is expected from our software capabilities. And we have a very strong team at Innoviz developing it. It's less easy to identify. I think that we're showing the product, the LiDAR, and I think that the perception software is much more difficult to demonstrate, and possibly that's why people are less aware of it. But a big part of what Innoviz is developing is the perception software. And I actually believe that in 3 to 4 years the majority of the company would become much more software oriented rather than just the rigs and wires. So I think it's one of the qualities that we have here at Innoviz and we'll continue to grow.

I would like just like to add that Innoviz has seen -- has identified perception as a strategic goal since the beginning of the company. The first team that we recruited, we had a perception software engineer. And since then, this only grew. So we are actually probably the LiDAR company with the most extensive experience in perception.

Yes. For many years we've been the only one that's developing perception software for the LiDAR and that's something that some others have started just a couple of years ago. And we're very satisfied with the progress we've made and the maturity of the product and that's also the feedback that we're getting from the customers. So well done, all.

Yes. That's good answers. And maybe going back to the use of cash and tying that in. You hired 35 people this year. Is there plans for hiring as many people or more people next year or during this year I should say, last year?

We hired 35 people last quarter, that's correct. And Innoviz, over the past year, grew quite a bit. It depends -- as Omer mentioned before, it depends on the projects that we win. We are very -- controlling our budget very tightly, very efficiently, and growth will be expected if we see that programs and business is coming in.

Maybe I'll add in short, but Oren didn't have much time to show, but his organization is built in a very, I would say, smart way. We have the core engineering alongside the InnovizOne engineering and then you have the 2 programs supporting the 2 customers in going to launch, and then you have the InnovizTwo and building block -- and other teams that are now working to support different activities. So I think that we are built in a way that is, I would say, efficient. So we don't need to double our teams in order for supporting multiple programs. But as Eldar said, it's going to be managed through acquisition of new programs.

Congratulations again.

You have no further questions. Please continue.

Okay. So we are actually expecting very exciting times now. We see very -- a lot of progress also from the R&D but also from the business side. And we are looking forward to exciting year. I think it's going to be a great year for Innoviz. I have huge confidence in what we do, and I'm very proud of what we managed to achieve. And with that growing the company very successful. Thank you very much.

Thank you.