Sivers Semiconductors AB (publ) (SIVE) Earnings Call Transcript & Summary

Hello, and very welcome to this Capital Markets Day with Sivers Semiconductors. My name is Anders Storm, and I am the Group CEO of Sivers. I'm going to take you through a very exciting agenda we have. Unfortunately, we don't have the moderator here today, but I'm going to moderate today. So we're going to take you through this. We have a very exciting agenda. I'm going to start talking about overall the company, then we're going to have Sivers Photonics. William, also called Billy, doing his presentation and then Sivers Wireless, which I will do. Then we're going to have the CEO of Cambium Networks, a customer to Sivers Wireless doing his presentation. And then we're going into Photonics and the partnerships with Imec and a very interesting deep dive into the Photonics business. And here, it's going to be very technical. So for those guys who likes that, that's going to be very interesting. And then in the end, we're going to end with a Q&A. So we're not going to have a Q&A during the presentation. So please send your questions into the app or we have people here as well that can ask questions straight to us in the microphone. So let's start with the video. [Presentation]

Okay. So that is a summary of what Sivers Semiconductors are doing. But today, we're going to tell you more about the details. So thank you for joining us today. So to go back a bit and give you sort of the introduction to the company, if you haven't heard about us before, we are a company working in 2 business areas. One is the Wireless business area where we're focusing on 5G technology today. The other one is sort of an area which is called Photonics, where we do laser chipsets. The company today and the group has had its headquarters here in Stockholm in Kista. And we are about 120 employees, and we also have an office in Glasgow and 1 in Gothenburg with the focus of Wireless in Sweden and the Photonics in Scotland. The company was actually founded quite far back. About 6, 7 years ago, we started a journey to go into these very exciting business areas. And I would say both the 5G and the Photonics business areas are now sort of really, really hot areas to be working in. So that's fantastic that we have products in both areas and making great progress. We also have a very good team behind this working on delivering all of these things from over 20 PhDs, who makes the products together with others as well. We have also strong investors. We brought in a lot of institutional investors, over 20% of the company now. We also have a strong cash position. We had $60 million in the end of Q2. So what has happened since we started this journey then and what have we done and where are we today? So if we go back down to 2016, we were listed on AktieTorget or Spotlight at that point. And we were contemplating how we could take sort of all this high frequency technology that Sivers have been doing for many, many years in radars, in voltage control, oscillators and so forth into this new interesting 5G market where high frequency is going to be sort of the most important thing. So we started that journey and started developing some circuits, RFICs and antennas. Also in 2017, we had the opportunity to acquire a Photonics company at that point, CST Global. We acquired them for $23 million, which has been a very good acquisition, as you can see. Also, we decided to sort of rebrand ourselves and get into the NASDAQ First North, which has a much stronger brand all over the world, and that was very important for us, of course, because we are working more or less outside of Sweden, most of the time rather than in the rest of the world. And we also, at 2017, already got our first design win in the 5G even before we had the first chipset actually ready. And now we've done sort of 26 design wins. And a design win means that you actually have a customer who have picked your chip and started to design in any way to making a product. So from that, and in 2018, when this product was sort of more or less ready, we actually won a very prestigious prize at the IEEE event in Philadelphia before a lot of the really big semiconductor companies out there. And from that, it only -- it took off really strongly, and we added more and more design wins over the years. We also added in October 2018, our first Fortune 100 customer in Photonics, which was a sort of massive step for us, getting an agreement sort of an MSA as it's called, getting that into place. And that customer have now placed sort of prefunding or NRE, over $10 million into the company in new products we're working with. In 2019, we kept progressing 5G and the Fortune 100 customers. In 2020, we also did the same thing. We added a new Fortune 100 customer as well to the list. We were up to 21 design wins in total. And we also had some 2 really large customers in the 5G where we could go out with volumes and so forth. So over $60 million sort of in estimated orders came in, and we also changed our name to Sivers semiconductors and rebranded the company from our previous name Sivers IMA. So Sivers Semiconductors is actually telling much more the story where we are as a company now than Sivers IMA, and it's much easier actually to be out there and market the company. People immediately understands who we are when we address ourselves as a semiconductor company. And now in 2021, we took the huge leap of moving into the main listing on NASDAQ. We moved directly up to the mid-cap list in -- on the 10th of June, which was a big leap. We also seen the number of shareholders growing heavily. We had almost up to 20,000 shareholders now. We've also gone over 20% in institutional investors. We have launched a new product that I'm going to talk about just yesterday, and we're also hiring a lot of people now to take care of the future growth. So we actually added more than 20% of the head count so far and are still adding during the year. So to mention some of the significant events, I mean we have added 5 design wins. We have up 26 now this year. We have 2 new orders from our Fortune 100 customer at SEK 25 million total. We got our first 5G volume orders from the U.S. at about SEK 7 million. We also were approved for trading in NASDAQ again. We had an order and a confirmation from 8devices that the technology is really working well. We also have joined forces with Imec and ASM for hybrid integration of silicon photonics, and we're going to tell you much more about that today. That's really cool technology, but also really hard to understand maybe. We also have our lead Japanese customer now moving towards important steps when it comes to mass production. We're establishing our U.S. office. We just hired the first VP Business Development in the Photonics business. And we also launched this really state-of-the-art 5G chipset just yesterday, and I'm going to talk more about. And that's many first in the market that we have been able to achieve with that chipset. Some numbers as well. Looking back at the second quarter, we actually had a growth on the revenue from 25% year-on-year. If we look at the wireless business, it actually grew 118%, which is good. We had some sort of lower EBITDA, but a lot of it or most of it is actually connected to onetime costs incurred because of the NASDAQ move. So in general, and assets we have in the end that's SEK 143 million, so -- cash. So we have a lot of cash left to keep on investing in the growth that the company is seeing. And here, you can see the segmentation reporting. As I said, the Wireless business grew really well. Photonics was sort of flat for the quarter. Our biggest market areas are North America with almost 70% this quarter and Europe 18% and Asia 13%. And of course, the reason for the North American market is connected both to the Fortune 100 customers and also that the U.S. market is first with millimeter wave, and it's coming now in different steps which I will talk more about in the Wireless business later on. If we look back how the financial performance has been in top line over the years. So in 2016, we had SEK 18 million. We've moved up now so we had SEK 96 million in 2019. Then unfortunately, we got straight into pandemic, but we're still able to reach the same level in 2020. So why are we here? And what's the sort of fundamental reasoning behind sort of what Sivers is doing? Yes, number 1, I mean, there is an increased data capacity need in the whole world, and everybody is probably aware of this and the market has been growing, and Ericsson report is constantly saying that it's an exponential growth. And we're using more and more data. We're using data centers, cloud storage, sending movies through -- watching Netflix. All of this is driving a huge need out there. And there's 2 technologies that can actually sort of provide the gigabit society in the future. And that is sort of the 5G piece and the fiber-to-the-home or fiber-in-the-data center. So all of these things is really important, and we are in the middle of this as a company and the strategy is also built on this foundation. And if we look at those different markets, I mean hyperscale data centers, which is behind all the sort of cloud data centers is growing a lot. There's a lot of different other areas within Silicon Photonics that's growing, health care is a new market that's growing a lot in there where you're using lasers for different sort of sensors and so forth. So it's a very interesting market as well in this area where we can address a lot of different verticals and so forth. And we have a fantastic sort of market in front of us with great organic growth. So if you look at where we are maybe strongest in 5G, which is the fixed wireless access piece, there is an estimate of sort of 88% growth CAGR from 2021 to 2025. If we look at the Photonics market, we had the health care sector growing at 81%, and we look at the sort of data centers about 20%. However, there's also a lot of other new verticals emerging in these markets. And how can you sort of get into those without sort of too much work in that sense and the long lead times? And what kind of market is it? I mean if we look at today in 5G, we're addressing a market around $7 billion TAM that we can sort of address. If we look at the overall market that we don't address that will use sort of these high frequencies, millimeter wave type of things, you have defense markets, you have handsets, PCs, tablets. You have SATCOM radar, 5G repeaters. There's a lot of markets here you can sort of merge into and reduce what we're doing or you can actually acquire companies maybe who can bring you into this very quickly. And there's a 10x market that we could address with this. So what we're saying it's -- there's possibility also to acquire growth here in very, very interesting verticals. So I mean, there could be 5G handsets, I mentioned, defense, SATCOM, radar or consolidation, of course, in the market we are already. Photonics itself, you can do vertical integration, go up in the value chain. There's consolidation. You need talent. You need capacity. For example, the Fortune 100 customer needs capacity. So there is many opportunities for the company here to grow and move in that direction. And to look back, I mean, we did a really good acquisition of CST Global. We look at the semiconductors in general, it is a very, very sort of busy market when it comes to acquisitions. It's been almost 1,600 acquisition over the last 11 years. So one important piece of the strategy, not just sort of the own growth and the organic growth is, of course, to constantly look at this and pick companies that fits into your portfolio and that kind of so. So what we're doing and what we've done so far really strongly, we've been working really, really hard with our partners to develop our products and have sort of companies around us that fits into the value chain. So we have partners like IDT, Renesas, NXP, Imec, who's here today, Ampleon, Blu Wireless, ASM AMICRA and so forth. All of those things is sort of very important. As you will see today, when we go into the details about Photonics, how sort of important it is to work together on these pieces. Also adding sort of the organic verticals and growing them, and we are in a very good position now to use all the products. We have a lot of products out there as well. And then in the end, sort of evaluate new high-growth verticals in M&A activities, both in Wireless and Photonics. I think that's an important piece to have in the overall strategy for the company going forward as well. So these are the fundamental pillars, what we are seeing in the future that could grow the company in a good direction. If we look a bit on the shares and the owners over time, here, you can see the shares since 2016 to now. There's been some sort of huge event here. For example, when we entered the First North and we went into the main list, the big order last year with the SEK 480 million order, for example. We have had a big change, I would say, over these years in the owner list with a lot of new owners, Swedbank, in total, has about 10-plus percent now. We've also recently got the third AP fund as an investor and so forth. We have a German fund now with plus 3% in the company as well. So I think we have a very good mix of owners as well as a lot of retail investors that is interested in the company. So it's a very good mix now, and we keep on building on that, of course, going forward. So I've actually gone through this a bit quicker, and I don't know if the guys are on line yet to -- yes, so they are online. So we will now move into Photonics -- and I'm happy to present Billy, who is our MD for Photonics, and he will now take over.

Okay. Anders, just check, you can hear me okay?

I hear you very well. Yes.

Okay. Good. Good. Okay. And as I can't hear you, but I think you can hear me. Okay, that's good. Thanks, Anders, for the introduction and hi, everyone, to an update on Sivers Photonics business and what we are doing. My name is William McLaughlin, as Anders said. And I'm the Managing Director for Sivers Photonics. I studied electronics and engineering at Glasgow University, space mining in photonics in the silicon industry. So I've got more than 20 years' experience in semiconductor development, quality and operations. And today, I'm going to give you an introduction to the Sivers business. The next slide. We're based in Blantyre where we have a semiconductor fabrication facility for development and manufacturing of photonic devices, specifically lasers. Our team is currently around 76 people, and we're expanding. We expanded actually this year by about 10 people. So we've gone from 70 to 80, and we've got plans to expand that even further significantly over the next year and the year after. Currently, that group consists about 14 PhDs, given a lot of our work involves research and development where we then take that and develop it into engineering and production solutions. We operate 6 days a week. So fabless is fully functioning over 3 shifts. So it's a full semiconductor fabrication facility. Within this facility and with our current capable team, we have developed excellent technical capability in laser photonics, which allows us to participate in multiple high-growth areas, and I'm going to explain that later in the update. Next slide, Anders. Here, I've got some examples of what some of the end state-of-art facility can look like. We have processes from lithography to etching. So it's a lot of material science, a lot of testing. And this is a type of capability that's required in order to fabricate the basis. We actually run our operations in 4-inch, which is actually a state-of-the-art for indium phosphide. And we have got plans to move that to 6-inch, which would be up to the kind of the latest has been done in the industry. So excellent capability exists that we can do end to end in our facility. And next slide, Anders. I'm going to talk now about the 3 main kind of vertical areas that we have other areas that we're working on and I'll talk about, such as kind of quantum in research and development. But the 3 main commercial pillars that we're looking at are optical communications, which is going through a kind of 2x growth at the moment up to 24, and it's expected to grow even further now. 4x in optical sensing. Optical sensing is where a huge win of our work is right now. It touches on LiDAR, biosensing and areas like that. And then there's optical wireless applications like -- sorry, LiFi which is coming in. So that's also expected to get through an enormous growth up to 10x growth. So we're involved in all of these. We're also working with several Fortune 100 U.S. companies in these areas, and I'm now going to briefly discuss a bit more about it. The first vertical is the optical communications market. This market varies from cloud data center products to quantum key distribution development. At Silicon Photonics -- sorry, Sivers Photonics are actively engaged in all of these areas, working mainly directly with customers, but also several academic collaborations. And we've emerged out of a lot of those academic collaborations going back over the years. A key markets for us, as I mentioned already, is in the U.S., approximately 90% -- 85% area of our business is in the U.S. and Europe. And we're currently working on product qualifications, which will then lead up to production ramps with several Fortune 100 companies in this year. Moving then on to the next area, which is optical sensing in sensors, a new and very exciting area within Photonics. Traditionally, Photonics has been thought always basic optical communications only. But now that is going into every area that's going to touch our daily lives. So this includes areas such as autonomous vehicles, consumer biometrics and augmented and merged reality. And we're seeing a lot of activity and requests in this from Silicon Valley U.S. companies with most of the applications and augmented reality and security, such as facial recognition and atomic clocks. The key technology which enables most of this to happen is silicon photonics, and this is a really exciting area we're going to spend quite a bit of this talk. And Andy and Joris are also going to talk with us later on in a little more detail. But we are developing significant expertise within Sivers Photonics in that area. So I'm now going to discuss silicon photonics in a bit more detail. So on the next slide. Just to cover the third vertical before I move on, this is still involved in an emerging market, where we're expanding our customers. So I would say of the 3 verticals right now, this is the area that we are probably working the least in. And that's actually driven because -- happening solely because of the kind of sensing in sensors area is so aggressive right now, and there's so much work in that area, but that's where most of our activities been spent. I mentioned earlier on that we also have quite a lot of research programs in [indiscernible] as in quantum technology. And we've got research projects and there from Quantum Atomic Magnetometer to quantum-based LiDAR and quantum key distribution. And the growth area for that has been well communicated is huge. Again, that's still very much in the research and development phase. So I'm now going to explain really what silicon photonics is all about. The graph that kind of I'm showing here is just kind of traditional. On the left-hand side, we have the classic kind of fiberoptic business, which is in nature long hauls, huge distances. And then we have the kind of silicon business on the right-hand side. The silicon business has gone through enormous changes only about 7 to 8 years ago, 28-nanometer was the most advanced technology. Now you have companies down to 3- and 5--nanometer and even lower than that like carbon nano chips and things like that. However, the real limiting factor there is the requirement in terms of power and RC delays, which are caused by copper and the insulators to get data in and out all the device, and this is where Photonics comes to that. So that's is basically silicon photonics, both meeting each other where we get the best of both worlds. So the data will be kind of brought in and out of the device using photons rather than electrodes, and this will be a game changer for the semiconductor and all the applications involved will be a game changer for the whole of the industry. To -- how can serious everyone's taken that, the number of companies that are involved in this. This map shows that. Some of the companies that are involved in silicon photonics and everyone who's involved in semiconductors or silicon right now is starting to gear up to this transition that's going to happen. And as I said, it will be a game changer. The applications will utilize silicon photonics and LiDAR, optical comms, biosensing, applications such as artificial intelligence and every space that you can think of, including quantum. Next slide, Anders. So looking forward, Photonics, seems a very attractive growth area, as I mentioned, because of the low cost and high volume semiconductor industry that already exists. So the infrastructure in semiconductors is already there. What we need is how to basically manufacture photonic devices, integrate them, and Sivers Photonics has the capability to actually meet that requirement. This technology is expected to reach SEK 3 billion in 5 years in areas such as biosensors for watches and the adoption of the smartphones and various other wearable devices. Silicon Photonics is also is one of the only -- sorry, Sivers Photonics is one of the only companies that offers photonics application capability across the board. Moving on to the next slide, Anders. Okay. So what are Sivers Photonics' key strengths? And I've mentioned quite a few of them there. We offer a unique custom end-to-end service from design and modeling, prototyping to high-volume manufacturing, up to 100,000 devices on any of our subjects. It's a huge number of devices that we make for substrate. And at a facility, we are capable to actually manufacture in this current facility very high volumes of that. Our main focus is going to be on our indium phosphide laser platform, and that platform will be used as an enabler for silicon photonics. We've already sold more than 50 million devices in this field. The one other key part of this is what's called it's quite a long word, but it's CW-WDM, is what's called an MSA, master supplier agreement. And Sivers Photonics has been chosen as a key partner in order to create basically the standard platform. So working with various other companies as a founding member, a promoting member on the standard. Andy and Joris is going to talk about this a bit more later on. But what that standard will do will allow all of the companies now a bit the way the silicon industry went allowing the easy adoption all the technology and to bring the application to the field much quicker. And as I mentioned, this will touch on everyone from artificial intelligence, optical comms and high-density co-packaging optics. We already have several commercial orders as a result of this. So it's very exciting. Next slide, Anders. And as part of this, we've already announced, we've got collaboration with Imec that Anders mentioned. Again, Andy and Joris will cover this in more detail later. We plan also to make further announcements as part of the joint development program. Okay. So as I discussed earlier on, going back to the kind of Fortune 100 companies or Tier 1 companies that we're working with, we've got continuous business with several. And I'm going to talk about roughly, I can't mention any details, but 2 companies that we've got a lot of activity with, and we are gearing up -- right now, we have activity started with others. So several of the key projects are tracked internally in a lot of detail. These are going to lead to high-volume businesses. So we're very, very excited about that we've been hiring more staff, prudent framework orders in place and starting to get ready for pilot reduction and high-volume manufacturing. Next slide. One of the other Tier 1 companies we're working with is, and as I mentioned already, is an augmented and merged reality. The Sivers are working on this area, and we're looking forward to further announcements over the next year. To add some detail to that. Would you go to the next slide, Anders? Some of the kind of detail that sits behind that. It typically takes about 4 years for that to come from initial discussions through to fruition, and that's because of the kind of long gestation period during the development phase and then into product development after the research debt. And we've already actually attracted quite a lot of revenue from those customers, and we expect to increase that much further. Okay. So in summary, again, we've still got a lot of detail that we're going to cover here later in the talks. Andy and Joris are going to go into a lot of detail on silicon photonics. I expect everyone to have a knowledge of that after today. Sivers photonics, what do we offer a unique technology, capability where we can design prototype and manufacture high-volume semiconductor photonic devices. We have a world-class expertise in indium phosphide platform development and silicon photonics indication. We operate in multiple high-growth markets including optical comms, sensing and wireless. And we're working with same Fortune 100 silicon U.S. companies on multiple projects. So very, very exciting future ahead of us. And again, that was quite a quick talk but that is all the material at the moment, and I'll be happy to take any questions later on.

Thank you, Billy. So let's move on then into the Wireless piece here. And I'm going to present to you -- oh, sorry, I need to speak English. So let's move on from this. I'm going to present to you the next step here with Wireless. And first, I'm going to start on the sort of the basics a bit. What is 5G? And of course, it's the fifth-generation mobile networks. And that's where the sort of the promise of gigabit speed has always been. And also this low latency and latency is the time it takes to a signal to go back and forth in the network, which is really important for mostly maybe for gamers, but also for autonomous cars or vehicles that need to be connected and do really quick decisions and changes based on some AI or whatever it is that steers the vehicle or self-driving cars, for example, who get the information. So fifth generation will enable that through low latency and high speed. Then of course, there are a lot of different parts of the fifth-generation mobile network. You have what's called the low band, which is sort of, let's say, a glorified 4G network with low frequencies, very small bandwidth, but they get very far in there. You have the mid-band 5G technology, which is sort of sub-6 gigahertz technologies. They go a bit shorter, but you can get a bit more speed, and that's probably what you're all seeing here in Stockholm when you're using a network, but you get sort of 100 megabit per second. You don't really get the gigabit speeds. Then you have sort of what we call the real 5G, which is the millimeter wave 5G, which is above 24 gigahertz and upwards. And that's where you actually get the sort of wide, wide bandwidth and you can actually sort of push through data much better. And it has nothing to do with the millimeter with signals being smaller. It's all about that there is bandwidth up there that has not been used and it's available, and there's a capacity crunch, I would say, in the lower bands because there's a lot of different things there that are sort of competing for the frequencies. So -- and there are 2 pieces of this. That's what we call the sort of the licensed 5G NR, where NR is the new radio. And the license piece, which is 57 to 71 gigahertz, which is using more WiFi technology and license, which we're all using every day. So these are the sort of very interesting pieces, and that's where Sivers Semiconductors is working in the 5G field. And a bit, as I mentioned before, you have below 50 megabits per second in the low band. I think T-Mobile in the U.S. was very early out with that, for example. Mid-band has been used a lot here in Europe with the 5G, and that is what you can see here. I tested Telia here in town, I had 108 meg down, 46 up. So that's what you get basically on mid-band. And then the 24 gigahertz and upwards there, suddenly, you have 400 to 800 megahertz or over 2 gigahertz of bandwidth, and that's where you get gigabit speeds and upwards. So that's really interesting. And then a bit more why it is like that? If you look at this, all the different technologies that's been done before 5G came has been on these sub-6 gigahertz technologies. And that's where 2G, 4G, WiFi, even the 5G, I mentioned, Bluetooth, 3G, whatever, everything is fighting on this area, and they have this small channel with up to maybe 60 megahertz or something like that. Then you have this massive amount of licensed and unlicensed bandwidth for millimeter wave. And that is now what's coming in the second wave of 5G, let's say. And there has been a massive spectrum auctions in the U.S., where there was $7.6 billion both of the spectrum from Verizon and AT&T and others. There's been this unlicensed spectrum now coming in, in both the U.S., but in Europe, actually as late as January 2020, where we have a lot of applications like the track-to-train applications or what I assure you also around Trafalgar Square with different mesh applications and fixed wireless access where we're working. So this is actually a very sort of fantastic area to be in, and that's our sweet spot as well for the future. And then, of course, this is a process for the whole world to move into this. And frequencies are something that the sort of PTS or the FCC or whatever body it has sort of has the right to give out and people buying the licenses or they get licensed free spectrum. And that goes slowly, I would say, in general, in the U.S., they always first with everything. They've been leading since sort of Android and Apple took over from earlier technologies. And then Asia, like South Korea, Japan and other countries has always been very, very early as well. And Europe has started, unfortunately, to be a bit behind on this. But we are getting sort of sub-6 gigahertz 5G now in Europe, and we're seeing some countries like Italy and U.K., I saw something coming out now and Finland have done something, but that's going to come a bit behind. But it is very interesting, though, to see that it's now starting to happen, and we have customers starting rollout. And you will hear, Cambium CEO today tell you more about how they're actually rolling out real products in the field right now about these things. So it is very exciting to see that. It's now happening and the bodies are actually giving licenses out. So what are we doing in 5G? What technology are we using? And what are we actually building? So we are using something called silicon germanium, and we're using RF SOI. There are 2 technologies for different chipsets. We make sort of a very highly great integrated circuits, which call the RFIC or a beamformer IC. And we also integrate them together with the antenna to sort of create this high-frequency product that is sort of necessary to send these things out into the air. And it's very difficult to do on these high frequencies. It's maybe 10x more difficult than do a sub-6 price. You need very specific competence, and that's what Sivers Semiconductors has with our very long history in this and also integrating these things because there is so much things happening nowadays with these kind of things because this technology needs to be beam-steered because it goes shorter. So you have to sort of send the power in certain direction, which called beam steering, and that beam steering can then create a very good sort of strength in a certain direction. And therefore, you use a lot of sort of channels and you build 16 channels we have on the products today. And yesterday, we launched a 32-channel circuit that no one has. So this is very important because the more channels you have, the more power you can get, and it's a sort of 50% more power actually with doubling the channels. So that's quite interesting. Then we use what's called evaluation kits to get our customers to integrate this. And also to make it easier for our customers, we've partnering up with a lot of different partners who makes the baseband or the signal who can be sent out into the air. And that is like Renesas, IDT, NXP and others we're working with to be able to do this. And of course, the ODM themselves that have their own baseband to integrate it into them. And then there is a number of applications now from the start, which is really interesting for this high-frequency 5G. It's the fixed wireless access. Actually, sort of wireless broadband to the home, 5G in general then, where we now see sort of some build-out in base stations where you actually connect to handsets as well. We are not doing the handsets today. Mesh and backhaul where you sort of send these signals back to the Internet on the network, which is an important part because if you have high speeds in the front, you need to have high speeds in the back so you can do it with millimeter wave or fiber. We have track-to-train applications, which means that you can actually have a wireless broadband to the train. I mean the best thing would be if you can connect the fiber to the train, but I won't work. So you can get the fiber in the air with this instead. So that's very interesting, and we've seen the first implementations of this in the U.K., for example. Then we have Vehicle to X application, and we've seen cars and stuff being connected with this in test ranges. Well, we have McLaren that has our stuff on the top of them. We have our customer in Fujikura in Japan just finished a big trial in Japan on buses with autonomous buses trying this out. So all of those 26 design wins we have now are moving now into the next step here. And this is very exciting to see that all of this technology we now developed over these years in this sort of fundamental silicon technologies are now moving into real and hardware and products. And these are the hardware and products we are selling, and they look exactly like this, and this is what is integrated into our customers' products. So we have chipsets and antennas, and we have different type of antennas that can steer up and down and vertically and horizontally. We have antennas that just sort of put the beam somewhere and so forth. So it's a bit different how you do this. We also have baseband partners today, NXP and IDT. We're working on more of them. And now with the new chip, I think we're going to be able to bring in more partners in that side as well. We also have a partnership with Ampleon, which doing a sort of more a beamformer type chip which is less integrated and is more used sort of in base station type of application. Yes, to mention that as well, I mean, all of these things, in general, we developed are really highly integrated. So it's really good to use in the sort of consumer near things like the home units, the CPEs or the small cells or the pico cells and that kind of things because it gives you less parts, lower cost and higher integration basically. So we have those 26 design wins. I'm going to go in to tell you more about them. I mean one of the bigger ones is the ADTRAN and CCS. We have Blu Wireless with the track-to-train applications. We have Cambium Networks, which will be here today. Of course, Fujikura, we just had a press release that they are now going into volume production. We have 8devices. We have partnered with Ampleon for the beamformer and so forth and so forth. And the really big project we're working on now is the CPE deal, and that's sort of the new product that we launched yesterday. They will go into this in Q1 2023. So we already need to have hardware now to be able to integrate that and that got into volume production than in Q1 with an undisclosed partner so far, but it's for the U.S. market. We have other customers like Siemens Healthineers, who use this for health applications as well. So there's a lot of ways you can actually use this technology. And even if you use it in different verticals, we don't need to sort of change the product too much. We can actually sort of reuse what we have. It's sort of a COTS product from the shelf in that sense. So it doesn't require too many people to do all of these things either. So to give you a more flavor on where the products are moving and so forth, and I'm sure that Atul will tell you more about this product that they are now launching and getting out into proof of concepts with their customers. And I'm not going to dwell too much on this, but this is a very exciting product we're doing with Cambium. Then we have Fujikura, as I mentioned, they built up what's called the CPE. They also made own antenna and a module that can be sold. And separately, they also have this sort of small base station, a small cell that they can use together for Vehicle-to-X communication and buses that it is in this case. And they're now going into volume production, and we're working on closing a deal with them now as well for the volume production. This is an extremely cold project also with Blu Wireless, where we have the track-to-train applications. And this is the first time you actually can do real gigabit connectivity to trains, and that's because of this bandwidth and the technology we're using that can follow the train at speed. So there's been tests done with these trains on 200 kilometers per hour where the train go passed and we can be connected in both the back and the front with 2 gigabits per second. So maybe we get 4 gigabits per second into the train and then you can have WiFi on the train, so everybody can happily watch Netflix between Glasgow and London or whatever you want on the train. And that's the next thing that's going to be built out with Blu Wireless and in this case, then First Rail, which is the owner of Evo-Rail have taken this product to market. And they are now talking to several large train companies. So I'm hoping within the coming 6 months, you will see more and more of these deals coming out. And actually, I haven't seen any sort of competitor at all to Evo-Rail here who can do this technology. Of course, this is sort of -- not sort of a big volume market per se like fixed wireless accesses, but it's a very good margin market and a high sales per unit market for us and it's excellent and cool technology, of course, to showcase what you can do with this in the future as well. ADTRAN, they have these mesh and fixed wireless access products. ADTRAN is a quite large company listed in NASDAQ in the U.S. They actually merged with another company now also, and they are actually a fiber company from the beginning. But they're going to use this technology as sort of what they call fiber extension. And they have a huge amount of Tier 1 operators where they're selling fiber solutions, too. So I think it's a very smart strategy from them to sort of prolong and have the sort of last mile or whatever connectivity where they can't put fiber and they can broaden this network out for any of their operators. And they're a big drive in the U.S., the infrastructure drive and the RDOF which is now giving money to operators to install gigabit technology everywhere. So we're waiting now with patience on seeing how well with ADTRAN be able to sort of get this technology out together with the fiber solutions. We also have CCS, who have built out this with sort of a neutral host operator in London called Ontix. So this is built out now in several places, and the network is growing every day and they're putting in orders. Of course, it's been hampered quite a lot as well. I didn't mention that with Blu Wireless, but of course, that's been delayed a lot due to the COVID situation, and no one has been on the train, so they haven't built it out. But this is the same thing a bit, but they built out some parts in Westminster recently. They have it in Soho now. They have it around Trafalgar Square as it's been before, and this is very interesting in general. And they are actually using it also to backhaul information from ordinary WiFi hotspot or that kind of thing. So it's fantastic to see our technology on lamp poles in London. I was there looking myself. It's really cool. Then this is a completely different technology where we're using Sivers technology inside on a company called AirVine. They call it wave tunnel. Wave tunnel is a way of sending gigabit speeds through walls and sort of replace cabling in house in buildings. And they are talking about 5 million buildings in the U.S. that needs to replace the 100-megabit cabling to 1 gigabit cabling. So that is, of course, a very interesting piece here. Will they do that by sort of taking a long time, pull out the cable, change all the cables, 2 gigabit cables? Or can they do it with this, which is much quicker and easier. It's kind of interesting to see how well they succeed. But it's a smaller company, but really cool tech. My guess is that Cisco would buy these kind of companies if they succeed, but we'll see. We have this fixed wireless access technology as well. This is being launched now in the U.S. with Tachyon Networks. We have a company 8devices that sort of builds modules based on our stuff to integrate Renesas, baseband, our antenna, making module a mPCIe IE module that you can just directly connect into your solution and you can build up and you get even quicker to market. And they're also part of helping some larger fixed wireless access companies in the U.S. to get this to the market. So this is a very interesting application and supports us. And now to the 5G NR chip we just released, and this is very exciting. The team has been working on this quite a long time now. This is part of the big CPE order we have talked about. And we have, together with this company, put down, and we always want to have that as we had with CCS on the 60 gig from the beginning. We always want to have an R&D company who sort of pulls this through. So we have a speaking partner on the technology and then also talk to the end market. So we actually know what to build into the technology. So they've been part of this and also sort of putting up the specification and all the needs and how we can build this in a good way. So this order includes that they are paying for development as well, and they will in future and pay per unit. So these chipsets have something that's never been seen before. We have 32 channels transmit and receive channels in total, which is fantastic, and that gives you sort of a very interesting solution. And you can then do vertical and horizontal polarization. So you only need -- it looks like you have 16 patches, but you actually have 2 polarizations below them. So you can make a really, really small module. Also, most of the chipsets competing out there has either zero-IF or low IF signal going in from the baseband. We have both here, which is really flexible for any customer. So we don't need to think about if they have whatever they can actually use it. And for example, NXP has been really keen on using the zero-IF part because it costs less if you have that kind of interface. So that is really interesting. Also, we're focused on getting a power solution out, which is reaching the limits what you're allowed to do in certain things. So for the CPE here, we can use 1 single chipset instead of several chipsets, which means actually it's the integration and the synchronization, which is a big sort of challenge for many, can be done very easily because it's on 1 chip, which is really good. Yes. And then there is a lot of technical things here. But I mean, 1 thing is that we're now covering the whole band. There is 5 different bands in millimeter wave for licensed 5G. They called n257 and so forth. We're covering all of those. So that's from 24 to 43.5 gigahertz. We have really small modules. Engineering samples for these will be available in Q4 and production samples next year in a year. And it's been a fantastic sort of inflow of new request when we launched this yesterday. So it's a very nice to be able now finally to talk about this officially. We have, of course, talked a lot with the current customer about it, but now we're getting to the next step here. And this is also what we're going to do with NXP to integrate this. We talked about earlier module, but we think that this module is actually the one we should do with NXP, and they have agreed to that. So this is sort of what we see now during the coming 6, 9 months, I would say that we can get to something with NXP. So if we compare a bit, without saying the name on the left hand here, I mean many of competitors here, they're using 4 modules to do the same as we can do with 1 because we have higher power and more channels. So we actually have about 3 DB extra, which is 50% more output power, even if you just have 1 module here. So this is quite interesting, and this is one of the largest companies, starting with the queue. So we are doing something very specific here, which is really interesting. And of course, that company is really big, and they can win a lot of our market shares and so forth, but all companies who don't use their baseband needs to go somewhere else to look for technology if they don't want to use them. And if they have their own baseband, they need to go somewhere. So this is a very interesting niche market we can address. And if we look at the sort of the main use case for these are small cells and CPEs. And looking at the just recent market report from small cell forum, there is going to be a lot of growth in millimeter wave here. You can see the sub-6 gigahertz is already out there, but the yellow pieces here is actually all the small cells that's going to be built out over the coming years. And here's the millions of small cells. And of course, in a small cell, we need to use more than 1 unit. It could be up to 4 or whatever. So that's really interesting. Then if we look at fixed wireless access, as I mentioned before, it's still the 88% present growth in the future. We're seeing being part of this from 2023 with this module. It's just 4% sort of the market. We will capture with this customer. So there's a lot of opportunities on top of what we have won so far in this. We can also see that EU or Europe is actually coming now from nothing in 2020 to small things in millimeter wave 5G NR, but it will then start developing over time here. If we look at then in the end here, we have 6G as well to look a bit forward. And of course, that's sort of maybe 10 years down the line. And what will 6G be about? It will also be about even higher frequencies. Now we talked about now 71 gigahertz. This is 95, 140 gigahertz is they're talking about and that will be settled during a lot of discussion, of course, in the future. And you're looking at microsecond latency. And we actually there already with Blu Wireless. We have 0.7 million microsecond latency with the track-to-train application, which is huge. Then of course, they want to have it even lower and they want terabit is the sort of discussion of speeds. But as they've been talking about now, they talked about 10 gig for 5G sub-6 and they're at 100. So that's always the marketing number. I think the real numbers would probably be around 100 gigabits per second in the end when we discuss sort of what 6G actually will do. And I would say the unlicensed 6G, as we see it, 802.11ay is already developing and could be sort of out earlier and that's where we see 20, 30, 40 gigabit in real speeds in the future and that's sort of a project we are looking at in the next phase of the high frequency on the license bands. So if we summarize Wireless, I mean, 5G comes in many different forms. You have to be aware of which form you're talking about. We are in the millimeter wave, and that's where you get the true gigabit speeds. We had an 118% growth in 5G in Q2, which is really encouraging. We're seeing now that the pandemic is moving away a bit. We have sort of ramping, and we have Sivers Inside in 6, 7 products ramping, and we're really happy to soon have our -- the CEO of Cambium tell you about that. We also have just launched now the state-of-the-art 5G NR thing that will accelerate us and the company from 2023 and forward. 6G, many companies are working about it. You will probably hear us do things in the future here as well. But what's interesting is the millimeter will be even more important. So this is not sort of -- I think that's just millimeter wave is just now and here and then it disappears. It's actually going to be more important also 10 years down the line. So there has to be a road map for the future. So we are actually way ahead, I think, of the schedule here, about half an hour. And I'm not sure because our dear friend Atul is actually based in California. He's not here now. So I don't know if we should take a small break. And maybe there's some coffee out there and reconvene in half an hour. So he has 9 hours, and it's going to be 6:30 over there. So I don't think he's sort of up and running this early. So let's wait for half an hour, and we'll back here 3:30. Thank you so much so long. [Break]

Welcome back to Sivers Semiconductor's Capital Markets Day. Now we're going to have the next exciting speaker and it's our -- is the CEO from Cambium Networks, Atul. I'm very happy to have you here. And I'm handing over to you to present your company and what you guys are doing.

Thanks, Anders. Good morning from Silicon Valley.

Here.

Yes. Thank you. So we are very excited to be here and thanks, Anders, for giving us an opportunity. I will introduce Cambium Networks, and I hope we get about 5 minutes of question and answers at the end, and thus thank you for -- move it to the next slide, please. So this is our safe harbor statement. I won't read it, but just in case anybody has any questions, I'm here, but that's our statement. Next slide, please. So now I'm going to introduce briefly to you the Cambium Networks overall. As we look at Cambium Networks, we are really focused on singular mission. Our mission is to connect the unconnected. And the way we do it is we have a multi-gigabit wireless fabric, which weaves together multiple networking standards and really demystifies, simplifies the different complexities and brings together a very high performance when affordable broadband connectivity from 300 meters to over 100 kilometers. And I'll explain how their solution works. But that wireless fabric is very multipurpose. It's designed so you can customize and create a purpose-built network, reliable network, resilient network. And Cambium Networks brings 5G, LTE, WiFi, all different technologies. And all of those solutions are managed from a single pane of glass, single cloud-based network management. We have a very strong belief that if it can be wireless, it will be wireless. And the multi-gigabit fabric is the new fiber. You really don't have to have digging going around. And in a very simple manner, you can install very high-performance networks. In Q2 2021, we achieved record revenues of $92.7 million and delivered 19.9% adjusted EBITDA margin. So that's good strong results. We have -- our history briefly is we actually were divested out of Motorola. So very strong DNA of RF and quality wireless products. We are headquartered in Schaumburg, in Illinois and have R&D presence in Chicago area, Silicon Valley, Bangalore. So these are some of our very key sites and Ashburton, U.K. So with that global footprint, we have a very diverse revenue mix across products and geographies. Our broad market presence with over 17,000 network operators worldwide and over 10,000 channel partners, we sell our products through 2-tier channel and very strong global channel distribution worldwide. This has created an attractive financial model. And we believe that our revenue and growth is very sustainable. And I will go through the key differentiation and how we -- what is our intellectual property, how we create that multi-gigabits wireless fabric, which is all managed from a single pane of glass. And this wireless fabric is very mission-critical. That's one of the key DNA of the company. While we use merchant silicon, different chips, we have very strong cloud-based RF algorithms, device-based RF algorithms. We deal with RF noise in a lot of developing countries where we install our products. We also provide a significant spectral efficiency. The number of bits per hertz per second, we can pump through our network. Our main solutions are, we have point-to-point products -- point-to-point products connect towers and buildings at a distance of over 100 kilometers if needed. And then we have point-to-multipoint products. They spread the connectivity. They -- in the last 10 kilometers, 15 kilometers, 20 kilometers, they distribute the bandwidth, buildings, campuses, residential areas, so that's point-to-multipoint. And then our WiFi solutions provide the high-speed bandwidth in the last 300 meters, both across indoor and outdoor spaces. Most of our radios are outdoor. And a very strong expertise in designing quality in tough climates from minus 50 Celsius to plus 50 Celsius, really tough, tough terrains. So now let's go to the next slide, please. This will show -- the next slide will show the wireless fabric. What I mean by the wireless fabric? Really, the 3 key areas of wireless fabric are, as I said, point-to-multipoint, point-to-point and the entire WiFi solutions. WiFi solutions include switching as well. So let me just briefly describe, if you look at some of the near-term next 12 months or so, what you will see from Cambium Networks, these are some of the new capabilities Cambium is releasing. Our 28 gigahertz 5G, where we are working closely with Anders and Sivers, that product is in proof of concept right now, and we'll volume ship next quarter in Q4. So that's 28 gigahertz 5G. And then WiFi 6E. WiFi 6E is a very exciting capability Cambium will be adding. There's about 1.2, 1.3 gigahertz of band being added in 6 gigahertz zone. So that next year will be WiFi 6E. And then we are adding a lot of capabilities in our product lines. There's a product called Medusa, which is a 14 x 14 massive MIMO product. I won't go into the technical details. But fundamentally, it gives you very high capability from Cambium. So some of the recent launches, which are giving the traction both in urban and the rural environments, we introduced 60 gigahertz millimeter wave, which is called cnWave 60 gigahertz, and I'll show some of the customer success stories there as well. And 60 gigahertz for us is a precursor to our 28 gigahertz product coming next quarter. And the 60 gigahertz product gives us good capability. If you look at WiFi goes, for example, in the last 300, 400 meters and then the cnWave 60 gigahertz takes over next 2 kilometers or so, and it has a meshing capability. so you can really mesh couple of nodes and go multiple kilometers and gives you -- it gives you a significant gigabit capability. And then when you go from 4 to 7 kilometers or so, the 28 gigahertz, 5G, fixed 5G product takes over. So that's what I mean by wireless fabric. And then all other wireless fabric is managed from the cloud management. It doesn't matter whether you have WiFi or LTE or 5G or switching, it doesn't matter. We weave all of that into that cloud management -- single cloud management. WiFi 6 is a major transition going on right now in the industry. Every 4 to 5 years, WiFi changes the architecture. So there is a change going from WiFi 5 to WiFi 6 right now. And Cambium has introduced very high-performance products there. And they have a software-defined radio. So you can customize different throughput frequencies, all of that gives you a lot of flexibility. And WiFi 6, as I said, gives us the last 300, 400 meters. And Cambium does outdoor in a very superior manner. So a lot of our WiFi products going to campuses, into hospitality, goes into the smart city projects, those areas. And then the fixed wireless broadband comes right behind and gives full connectivity from that edge all the way to the core. And then 28 gigahertz cnWave, that is purpose built, fixed 5G point-to-multipoint. What I mean with the purpose-built is 5G will have a lot of complexity in big 5G core products. We don't do mobility. We only do fixed 5G. So we -- it's very purpose-built. That's what we mean. We have made it very simple, very easy to install and MU-MIMO. So this way, Cambium will have affordability yet mission-criticalness in a lot of applications where customers really want standards-based fixed wireless broadband. It's a very, very attractive product. And there are over about 30 customers or so worldwide who are now waiting for the proof of concept and we'll start that next quarter. That is one of the key collaboration, very innovative collaboration between Cambium and Sivers for many years. And we are very excited to launch this product next quarter. Next slide, please. So as we go into Cambium infrastructure differentiators, really, it is focused on -- very focused on 4 or 5 key unique things we do. Number one, we have leading spectral efficiency. And the spectral efficiencies are algorithms, the way we deal with air algorithms in our software, the way we design our antennas, there's a lot of unique systems architecture we put together to give leading spectral efficiency. And as I mentioned earlier, we have the product called Medusa and Medusa fundamentally gives you that phenomenal massive MU-MIMO architecture. And we have been shipping the product for last many years and is viewed in the industry as one of the very high spectral efficiency products. And now as we introduce the 28 gigahertz 5G fixed direction, if we go in that direction, we also have broad channels and very high -- multi-gigahertz high performance. So the journey of spectral efficiency by Cambium continues very strongly. We have embedded network intelligence. What we mean by that is we have -- we are constantly sniffing the air. We are constantly making sure we understand the noise conditions, the conditions in the spectrum and we adjust continuously and that gives our products a lot of resiliency and high quality, reliability, as I said earlier. All our designs are designed for minus 50 Celsius to plus 50 Celsius, tough terrains, tough climates in many, many different locations. And reliability wise, we also make -- we have very good strong labs where we test these products very thoroughly before they go out. Very scalable architecture. The way our entire cloud management functions, the cloud management is scalable. We have over 600,000 devices now under management in our cnMaestro cloud-based management. Very scalable architecture in terms of radios, how we manage them and number of users we can support. That spectral efficiency also feeds into scalability. If you take Cambium Networks' product, per access point on the tower, we can have a lot more users compared to competition. In many cases, even twice because that's how we've designed the product and very attractive economics. The way we do attractive economics is we really focus a lot more on merchant silicon partnerships with companies like Sivers. And these partnerships are deep. We don't just come at the end. We work with our partners way in the beginning and as we're designing. And so our partnership with Sivers has been -- and that's very close. Our engineers have jointly worked on a lot of things, debugged a lot of things. And that creates -- and then add value in that management performance, our spectral efficiency. Those are the areas we add value and that gives a very attractive economics without sacrificing quality. So while we have high quality, high performance, we maintain the economics. Next slide, please. So as we look at the broad set of customer base, as I said, our revenue is really diverse. Our entire presence is very global. And this gives you in different segments, some of the customers we serve worldwide. As I said earlier, we have 17,000 network operators worldwide, pretty global footprint and about 10,000 plus channel partners, 2-tier distribution. And the segments we serve is mostly Cambium goes after midsized service providers. And internationally, we do have large service providers. But typically, Cambium's value proposition is very strong for the midsized growing entrepreneurial and progressive service providers worldwide. And as you can see, whether it's Asia or Europe or North America, South America, good strong presence. And then SME, mid-market enterprises. In the enterprise segment, where we have been able to grow very well and enterprise is one of our fastest-growing segment, we have hospitality. So COVID impacted everybody. So enterprise did slow down in 2020, but 2021, enterprise has come back, especially segments like hospitality, education, smart city, some health care, these are the segments where they are bringing more connectivity, more throughput and really bringing lots of institutions online with that high-performance connectivity. Government. Many of our projects are with defense or local and state government, emergency response. Those are areas which need mission-critical connectivity. I call it ad hoc connectivity, connectivity on demand. Wherever there is an action going on, they need to create a broadband network quickly so that all the people are connected who are responding to the event. So government, local and state, defense, very key point-to-point, point-to-multipoint and sometimes WiFi usage as well. And then industrials. Industrials for us are oil and gas. We are present in many large oil and gas companies, providing whether it's point-to-multipoint or point-to-point from the oil platforms to the shore. Mining is another area, energy grid, water and waste management, transportation, particularly railways. Many of the control signaling uses fixed wireless broadband. So these are some of the very attractive segments. They all value that mission-criticalness, yet affordability. Next slide, please. As we look at the last, I would say, 15, 16 months, the world has come to Cambium Networks. While wireless was very important always, and it doesn't matter what segment you are in, wireless is always important. I would say, the need for more broadband from home because learning is increased at home. Students are working from home. Knowledge workers are working from home. So in general, the need for more throughput, more connectivity from home has increased. Enterprise refresh cycle with WiFi 6, I think we talked briefly about, that is going on right now. And then 5G and next-gen wireless is enabling the thing I mentioned, which is wireless is the new fiber. I think when fixed 5G arrives and it gets going, you will see now in addition to 60 gigahertz type capabilities, we're increasing the distance and now you have broader channels. So the multi-gigabit capability gets enabled as 5G and next-generation wireless comes. And broadband proliferation in general, whether it's the people, places or things, smart cities, digitization, more sensors everywhere, that means you really need a lot more connectivity, a lot more throughput. So the wireless fabric weaves together all of this and brings that entire solution in such a manner that you can easily deploy, easily manage. Next slide, please. I won't go into the details of this slide, but I think this kind of shows you how that wireless fabric works. And the point-to-point is those blue dots. And the point-to-multipoint, all the yellow dots and the WiFi shows the last 300 meters or so. So fundamentally, the buildings are connected with point-to-point. And there, you can go, as I said, up to 100-kilometer line of sight. Typically, it will be 10, 20, 30-kilometer connectivity from regional office to the headquarters or connecting campuses or buildings. And we have those very high-performance, high-quality point-to-point products. And they use different frequencies. They use different technologies depending on what is that purpose-built network. And then point-to-multipoint distributes the bandwidth in the last 10, 20 kilometers. And again, multiple technologies. There's a proprietary technology. There is now 5G fixed emerging. We give customers lots of choices because each customer has different need. Each nation has different frequency and different standards sometimes. So we give that flexibility so you can build purpose-built network. And in the last 300, 400 meters, we have indoor-outdoor WiFi. This is kind of how that entire campus or entire small city is covered with affordable broadband. And then all of those products in this diagram, you can see are managed from that single pane of glass in the cloud called cnMaestro. LAN and WAN are converging because earlier I used to see LAN was high speed, WAN used to be slower speed. Not anymore. With the technologies like 28 gigahertz and then 60 gigahertz coming in the last few -- the millimeter wave push happening, you really have fiber speeds and multi-gigabits all the way in last 10 kilometers or so. And then with meshing, you can extend those distances. We are also striving now to monetize software because we have lots of devices under management. We are beginning to add value-added services. There's a product called cnMaestro X, where we are basically starting to now monetize some of the capabilities, which are value-add. And then the Tier 1 and Tier 2 service providers in general, fixed wireless broadband is accepted now as a standard. That's one thing 5G has done. 5G has legitimized a lot of fixed wireless broadband standards. And as a result, we are very excited as high-performance 5G products come. Next slide, please. And very briefly in this slide, what we are showing is this is that concept of wireless fabric. I'll work from the right-hand side to the left-hand side. So if you look at the right-hand side, you have home gateways, you have ePMP 4000. Those are devices on your buildings, campuses, bringing the bandwidth to the end -- at the end. And they generally give you good distance, 10 kilometer, 15 kilometers and the performance of those devices -- these are end devices in the buildings. That's in a 200 megabit, 300 megabit type performance. Then you go more to the 5 to 6 kilometer -- into more distances. And as you go to more distance, you'll really start to bring that 5G fixed, then we have a PMP point-to-multipoint 450 products. And they start to give you 400, 500, 600, 700 megabit per second. Then you come towards 60 gig products, which are the last few kilometers and you start to see pretty high performance end-to-end. Both 28 gigahertz and 60 gigahertz will give you multi-gigabit performance. And all of these products, we can connect through our switching portfolio as well. So this is that concept of fabric that you purpose-build depending on the throughput, latency, what kind of network you need. You select the right technology, we weave together the fabric and offer you a complete solution. Next slide, please. I will share some of the case studies with you. I think that will give you a flavor of how these solutions are used. This case study is from -- it's a hybrid fiber wireless infrastructure, which delivers multi-gigabit to the home in Anchorage and Fairbanks, Alaska. That is one of the key points I want to make. The solutions we are pulling together, they are pretty agnostic. Sometimes wireless service providers use it. But a lot of times, the fiber guys realize that it's much easier to extend the network and some terrains are also very tough. And they have the billing systems and management systems. So it's easy to use wireless. And that's what we're finding. The Alaska Communications is a good example. In this case, their challenge was they really needed quickly to add the extension to the infrastructure. And they used Cambium products. 60 gigahertz distribution node is the kind of main node, which provides the bandwidth. And then the edge node is V3000 and they have deployed very successfully these 60 gigahertz product and provided a multi-gigabit connectivity to the homes and businesses in a fairly fast time with the same thing they could have -- they would have taken a lot longer, a lot more time if they had just gone pure fiber. So that's an example. Next slide, please. Next, I will share with you a customer called Pentanet. They are in Perth, Australia. Perth, Australia, actually, at one point, they were -- they didn't have -- they were the second slowest Internet speeds in Australian capital cities. And very quickly -- this company actually used to cater to gamers. They always focused on high-performance network in the city. That's kind of how they started. But when they looked at Cambium products, especially 60 gigahertz, and now we're going 28 gigahertz, they're very interested in that solution as well. They said, this is the way to bring both in remote areas as well as some high-density areas, a high-performance, multi-gigabit connectivity. So they adopted millimeter wave. And certainly, the network is giving phenomenal performance, especially to the gamers' low latency. And we just did a customer and channel event called Cambium Connections earlier this week. And Stephen Cornish, the CEO, was one of the speakers. And he said very satisfied with these advanced new technologies as they're coming in, bringing that connectivity in an affordable manner. So Pentanet is a great example how millimeter wave is being used to bring that high-speed connectivity. Next slide, please. This is an example of a Silicon Valley city. And their focus was to bring connectivity, affordable connectivity, but very high-speed performance both in that city as well as some of the areas where there was affordable housing. So they want to make sure that as they connect some of the government buildings, some of the residential areas, there's a unified architecture. So they used 60 gigahertz cnWave solution here. They have connected many institutions and very happy as they're expanding. All these millimeter wave customers are also now waiting for the fixed 5G expansion. As I said, we are covering with the mesh architecture last few kilometers and with 28 gigahertz, that distance increases and the multi-gigabit millimeter wave really takes over. So that's it. That's another case study of one of the cities in Silicon Valley there. Another very briefly, the case study is the gigabit speeds, municipal broadband offered in the city of Colorado -- Aurora, Colorado. All of these examples, what you see is that each network is a little different. Each economics is a little different in the city. And they -- yet, what they focus on is how we provide the future, how we provide performances which can keep scaling. And that's where some of these millimeter wave technologies are clearly going in because many places, they can't provide fiber. Many places, it's -- another example we have actually publicly talked about was in cities which are heritage sites, where they don't want to digging. So what we are finding is that wireless is becoming a very, very key way in which both wired and wireless guys are adopting, the next-generation architectures. Next slide, please. This is, I think, my last slide, and then we can open Q&A as well. So when it comes to the millimeter wave, we -- especially the 28 gigahertz, our heritage as a company is how to design purpose-built network that are easy to deploy? That's how we started 10 years back with our first set of products. That philosophy of easy to deploy, easy to manage, using merchant silicon, having that cloud radio algorithms has continued. And this is what we work very closely with Sivers in this 28 gigahertz product, 5G and our fixed product, which is next quarter we'll ship. And then it will continue in terms of the evolution, continue standards-based. Performance is fantastic. As I said, we are doing proof of concepts right now. And I'm very proud of the way the innovation happened between the 2 companies, the collaboration happened between the 2 companies, how our engineers work together. And the entire solution has a very holistic view. Right from the beginning, we focused on configuration, deployment and 5G is a sophisticated algorithm. Since we don't do mobility, we focus on fixed solution, high-performance solution that makes it purpose-built. So we're very excited working closely with Sivers. And with that, Anders, if you want, we can open the Q&A.

Thank you, Atul. I was thinking that we would take the Q&A afterwards. So if you can stay for another half hour or do you need -- want the Q&A now that we talked about? I don't know if you can hear me actually because that was a problem before also with Billy. He couldn't hear what I was saying.

And Anders, I can stay as well. If it is at the end, that's not a problem for me.

Yes. So we'll take the Q&A afterwards. I don't know if you hear me actually, but someone maybe can tell you that off-line over there because we have some connection issues. But I hear you, and we now move on and we have the Q&A afterwards.

Okay. Thank you very much.

Okay. Thank you very much, Atul. Very, very interesting presentation, and we have some questions already coming in now, so we'll take them soon. But now we're going to move on, and I assume our next speakers are online already. I'm trying to look up here if we have that. So we're going to have 2 speakers who's going to connect together here. It's our CTO from the Photonics side, Andrew McKee, also called Andy. And we have from Imec, Joris Van Campenhout, who is the Program Director of Optical I/O. And can I get the thumbs up here from someone that they are okay? Yes, okay. So we're now going to start, and we have Andy McKee, the next speaker. Take it away, Andy.

Okay. Thank you, Anders. And thanks for the introduction. So I'm Andy McKee. I'm currently CTO at Sivers Photonics. And I was one of the co-founders of CST Global back in 2001. So I have a 20-year history of working in this business. Prior to that, I worked for Hewlett Packard for 5 years and I have [indiscernible]. My personal background is indium-phosphide laser chip designing. I also spent a lot of time in those of technical road map and strategy within the business here at Sivers. And on a day-to-day business, I spent a lot of time supporting new business developments for the company. So in this whole [ maybe ] section, we're going to take a look at the silicon photonic industry from more of a technical perspective than home care. So firstly, we wanted to go back and revisit the slide that Billy presented earlier and just add a few further comments. So the top graphic that we have really demonstrates the wide range of companies involved in adopting and deploying silicon photonic technologies. The initial silicon photonic market was really traditional fiber optic communications markets and specifically integrated silicon photonics markets were developed to drive data center transceiver margins of up to 100 gigabit database and deployed extensively from data centers since then. So more recently, we see many more sort of entrants into the space. For example, LiDAR for autonomous vehicle sensing, biometric sensing and also health care applications. So the whole ecosystem is now rapidly growing and really diversifying. So it's important to stress that Sivers Photonics is involved with commercial activities in all of these high-growth areas. Even I would say, it largely makes sense, so that's very positive for us. From a technology perspective, and this is quite interesting and important. We're actually making one of our indium-phosphide 100 platform extremely similar laser devices for each of the different end markets. So for example, we can make 100 milliwatts high-[indiscernible] DFB laser for optical communications modules for LiDAR systems, but also for gas and biometric sensing applications. So that really makes it very efficient for us to manufacture devices on the platform. Next slide, please, Anders. So in this slide, we sort of dive a bit deeper into specific technology and where it affects the rules of ecosystem. It's sort of worth pointing out at this point the reason the 3-5 devices are required is the [ integrated ] silicons in the light. And therefore, 3-5 materials are always required and will always be required to optically cover these circuits. So today, there's a number of alternative technologies that are used to integrate the 3-5 onto the silicon. And they include grating indium-phosphide directly on the silicon wafers and this is actually showing this on ultimately that segment on this page. But best technology is very immature despite more than 20 years of research. There's a couple of other approaches where pieces of unprocessed 3-5 hetero-epitaxy bond is on the silicon wafers. And they subsequently process [indiscernible]. These approaches both implementations in terms of reliability and also quite inefficient use of a 3-5 material. And so we operate specifically highlighted in red and so top left part of the diagram is really the -- I'd say the complete fabrication and the testing of indium-phosphide lasers are known good die before they're actually bonded down onto the silicon wafers. So that's using very high accuracy with flip-chip assembly tools to do that. And this is really the most mature area of the silicon industry. And that gives us a number of sort of a combination of improvements both on the 3-5 chips and on the silicon wafer processing. But I would say, mainly by improvements in the placement accuracy of these flip-chip assembly tools. I mean again, just to reemphasize, as I said in the previous slide, we're using very generic integration technology from 3-5 to the silicon at multiple different markets and application areas. Okay. Next slide, please, Anders. Okay, that just quickly highlights the area that we specifically participate in. So this next slide gives a little animation of how our 3-5 photonic devices are flip-chip assembled onto the silicon wafer. So you can almost imagine our smaller indium-phosphide laser chips as a piece of label effectively snapped into place with very high precision to an underlying substrate. And this is a really critical part of the process and now that controls the optical coupling efficiency between the laser and the silicon photonic wafers, i.e., that depends how much of the laser light gets launched into the silicon photonic circuit. So hopefully that gives a very nice representation of how our lasers are assembled into silicon photonics circuits. Okay. Next slide, please, Anders. Thank you. Okay now, so the key strengths that we basically have today are the following: so we have very advanced in-house laser chip design capability. Specifically, we've got very strong capability in designing high-yielding DFB lasers, which support the applications of optical sensing and optical comms that demand an array that's really critical. We have wafer process technology, which has been developed to optimize and facilitate mechanical and optical integration into the silicon photonic circuits. Some of the details of this are shown at the bottom of this page. And we also see a schematic representation of all our sort of 8x array of DFB lasers light look like. Now while we obviously have a very close partnership with Imec that we'll talk a bit more about later, it's important to mention that our devices are also currently fabricated and integrated into silicon photonics circuits that are fabricated by multiple silicon photonics fibers, such as GlobalFoundries,, BTT and [indiscernible]. So the process architecture that we have today has also been [indiscernible] for high-volume scalability. Many of the applications we're in today are potentially very, very high volumes in excess -- or the magnitude in excess of what we do today. And we've done that using some -- complete all different processing technology, including etched laser facets and also high-precision optical coatings. And finally, on the Photonics side of the business, it's really important to emphasize, we have a fab that's a really important strength today given the well-publicized squeeze in the semiconductor industry, supply chain in general. So having a fab under our control is really important for the Photonics business today. So the 100-millimeter wafer fab that we have today is very well established with the full end-to-end process capability with capacity to further expand our outputs. And we've invested heavily over the last 4 years and additional crews brought in to expand our capacity, our capability and generally strengthen -- our general manufacturing strength capability. I'd also comment that there's only a handful of other commercial fibers globally, which have this technical design of fabrication capability to produce, in particular, these very advanced indium-phosphide DFB laser arrays. Okay. Thank you. Anders. Next slide, please. Okay, just a slide to highlight our indium-phosphide manufacturing platform. I mean as the title states, our InP100 platform underpins all of the indium-phosphide device manufacturing done today at Sivers Photonics. So what are the key features of the platform? Well, we're processing 100-millimeter wafer sizes. Not many fabs are doing that today, particularly for indium-phosphide photonic devices. We can yield up to even 125,000 laser die sites per wafer at the 100-millimeter diameter. We have in-house Ebeam lithography for DFB gratings and that's really critical in terms of having the design control to make these very high-yielding arrays. We have on-wafer -- facet etching on-wafer optical coating, as mentioned previously. And we have optimized architecture for the flip-chip bonding and it's all scalable to high volume. So in terms of the devices that we catch a manufacturer on the platform, a broad range of devices, but mainly focused on the emitters, the lasers. Obviously, we have DFB lasers across a broad wavelength range from 1,250 nanometer to 2.35 micron. And that covers the main wavelength range -- ranges of interest for both the fiberoptic comms markets, but also the optical sensing applications. We can fabricate devices with modulation rates from CW. That's constantly switched on laser to very high modulation rates of 28 gigabits. The CW lasers that we manufacture today probably dominate, I would say, and those are all externally modulated generally on the silicon. Okay. And we produce devices with rated optical powers of up to 100 milliwatts and pushing towards sort of 200-milliwatt mark, and that's driven by the requirements of silicon photonic applications. So the devices are also designed in such a way that they can support a very broad operating temperature range from minus 50 C to plus 95 C, and that covers the entire sort of industrial laser requirement space. So we manufacture also reflective optical amplifiers and they're used in external cavity tuneable lasers. And again, we can supply devices in single and array output format, mostly an array format mode just for scalability and for general sort of high-speed density. So one of the key takeaways really from all of this, 3-5 silicon photonic technology really is going to transform the electronics and photonics industries in the following markets. We have AI and machine learning, we have optical computing, we've got high-density copackaged optics for communications, we've got more traditional optical communications, we have LiDAR, biosensing, health care and even quantum technologies. So on the platform, we're manufacturing a range of devices or a very sort of spectrum of markets and applications. Okay. Next slide, please, Anders. So I think this is effectively my final slide, and it's just a quick introduction really to the partnership that we have with Imec and ASM AMICRA. Of course, he also covered this in a bit more detail in his slide decks, but I just wanted to comment quickly on what we're ultimately trying to achieve in this partnership. And that's really to accelerate the adoption of silicon photonic ICs. And we need to provide a one-stop solution for our customers that covers both the design, the fabrication and also the packaging of these circuits. And already, as Billy mentioned previously, we're seeing strong commercial interest with the partnership that was already publicized back at OFC in the summertime. So with that, I'll finish off and hand over to Joris, who will take us through the Imec slides.

Thanks very much, Andy. So yes, my name is Joris Van Campenhout. I'm Program Director for Optical I/O Research Program at Imec. And I will need to talk a bit further about what Andy already alluded to the collaboration we have in integration of indium-phosphide light source with silicon photonics. I will put it in a bit of a wider perspective of the other work, that work we're doing at Imec and I will get back with some more details on the actual collaboration. Next slide, please. So for those that are not familiar with Imec, I've included 2 introduction slides here to just illustrate what Imec is all about. We are a world-leading independent R&D and innovation hub, active in Nano and Digital technologies. So we are mostly based out of Belgium, where our headquarters is based. But we do have a global presence. So we've multiple sites both in the U.S. and in Asia. And the typical companies we work with and for includes a wide variety, including big leading semiconductor players, so the IDMs and the fabless, but also the foundries for that matter as well as system companies. So we work with these kind of established companies, but also with start-ups and also with a whole set of universities and academic groups. So typically, we do co-development partnerships, but we also have a variety of research programs where we work in a pretty competitive moat with multiple parties on a next-generation technology development. Next slide, please. So here, you can see a quick picture of our facilities and our headquarters in Leuven, so that's 25 kilometers east of Brussels, where you can see our 2 main R&D fabs, the older one is a 200-meter based one. It's based on 90-nanometer 0.13 micron CMOS. This is a facility where we have been developing silicon photonics at the start of our project, let's say. That switch is already more than 10 years ago. In the meantime, we have, let's say -- I would say, state-of-the-art fully functioning silicon photonics platform operating there. So we do quite some prototyping work with customers. But we also have, in the meantime, established a path to high-volume manufacturing with -- in a partnership with a commercial foundry. On the more advanced R&D sites, we have our 300-meter facility. There we, in fact, use more advanced tool sets. That's the area where we also have the research programs running on next-generation logic technology pathfinding as well as memory programs and advanced packaging. And so we are very lucky that we can also use that facility for advanced silicon photonics R&D. Next slide, please. So in the next couple of slides, I will quickly touch upon the main application drivers. Andrew already alluded to some of them. And I will recap that a little bit here. So next slide, please. So yes, the first driver that has originally, let's say, started most of the work or development work that we have been doing in the field of silicon photonics have been driven by optical interconnects. That's, of course, driven by the cloud systems that I'm sure everybody is very familiar with. So these cloud data centers, especially the hyperscale data centers, they are very big facilities that span several hundreds of meters in cross section, and they require a lot of optical interconnectivity. So the likes of Google, Microsoft, Facebook, they really have been driving the fields. I would say, since 2016. The 100-gig Ethernet era has really kind of driven very strong deployment of silicon photonics-based transceivers in this industry. And so that trend is continuing. We see very strong growth in this field driven by pluggable optics. But more recently, also the sheer bandwidth density requirements and power envelopes required the switches in such network, require a more advanced ways of integration of these transceivers with the advanced logic switch systems. And so on the right, you can see here a picture that's broad effect from a Cisco presentation where you can see that transceivers are now also being proposed to be co-integration in the first level package together with the Switch IC at capacities that spend 50, 100s and perhaps in the future also 200 terabit per second of aggregate bandwidth capacity. These are already mind-boggling numbers. And Imec is really active in the field of doing the technology pathfinding and development with these type of players to develop the silicon photonics technology that will enable that scaling in the foreseeable future. Next slide, please. Another area where we see a very strong demand for interconnectivity is the space of AI, machine learning and HPC systems. So you know that these kind of systems that typically use a multitude of GPU or TPU computing nodes, they also require terabit-scale interconnectivity between these compute nodes. And as these system scale outs and require more bandwidth, there is also a transition foreseen from going from copper interconnects to photonic interconnects perhaps in the next 2 or 3 generations of this type of technology. And so we can see that some of the leading players in this field that their research teams are already actively proposing new architectures that make use of these co-packaged optics where every single GPU would also have an optical transceiver, again, running at several terabits per second and high bandwidth density. And so that, of course, is a field that's seeing very strong growth as well and that can drive, I think, will be a next wave of further growth in silicon photonics transceiver technology. As you may notice here on this schematic there on the right, these brownish kind of modules, these are supposedly the lagged sources. In this case, multiwavelength light sources that would need to feed all these transceivers that are copackaged with these GPUs and other systems. So that's again, driving the need for more advanced lower power, higher bandwidth density silicon photonics technology, but also accompanying a light source that also needs to scale in complexity, for instance, number of wavelengths served. Next slide, please. So beyond optical interconnect applications, Andy have already alluded to it as well, we see more emerging sensing-based use cases of silicon photonics. So LiDAR and 3D sensing at large is one of them. So today, it remains a hard problem. But yes, basically, we're looking for a LiDAR solution that can be scaled down in costs, also in footprints. And today, systems are typically too big to get to these specifications. So doing chip scale integration of, for instance, the beam steering elements and meters and the receivers is something that we're also actively pursuing and researching in Imec. And that's also an area where more advanced light source for CW LiDAR may be required in the future. And in the next slide, the final application I would like to touch upon also in the field of biomedical and biosensing, also spectroscopic sensing applications. We see some kind of a similar situation where current systems are still fairly bulky. You can see a few example of molecular diagnostics, DNA sequencing, cellular analysis. These are fairly bulky systems and at Imec, we also have a number of research programs looking at miniaturizing these kind of systems down to the chip level, which, if you're successful, will enable, I would say, many more use cases of these kind of technologies. Next slide, please. So what I can do now is walk you through, let's say, some of the technology platforms we have been developing to address some of these applications that I just introduced. So in this slide, you can pretty much see, let's say, probably our most comprehensive, integrated silicon photonics platform that we have built, which is mainly tailored towards the communication use case, so this data center interconnectivity and also the emerging HPC AI system interconnect space. So we have, over the past 10 years or so have been quite active and successful in developing all these high-speed modulators. So you have a whole variety of them, the Mach-Zehnder swing of layers, electro-absorption modulators, using silicon and germanium at wafer scale. So we have this platform available in our 200-meter process line, but also first version in our 300-meter process line. We have our silicon wave guys there that are patterns using advanced [indiscernible] nanometer lithography. We have silicon nitride waveguides, we have high-speed detectors and we have the features to couple with low losses to optical fiber. But what you see missing here, of course, is a native light source solution. And again, we need here hybrid integration formats, and we also introduced that there are several ways of doing that. But the most mature one and the one that's most ready for near-term adoption in products is for sure the flip-chip integration and that's exactly what we're developing with our colleagues at Sivers with AMICRA. If we go to the next page, you can see here an example of such a transceiver prototype that we published at least 2 years ago. This was kind of a test chip that came out of our research program where we have, in this case, a 300-meter silicon photonics wafer that we fabricated and that contains a test chip for this terabit scale transceiver functionality. It also contains a flip-chip singles die here. So we're using also the most advanced 3D assembly techniques, including microbumps as you will see in the next image. If you can just click once more. This chip -- this interposer, in fact, is a [indiscernible] silicon photonics interposer that contains all the active high-speed optical components, the fiber coupling structures, but also through silicon vias. So it's really containing the most advanced 3D modules along with the most advanced optical components to get down to the terabit per second per millimeter bandwidth density. So very compact. It uses multiple wavelengths. It's not shown here in the slide, but that's a key feature of this technology as well. And so it's really kind of the technology that now several companies are starting to commercialize for this next-generation optical interconnectivity beyond, let's say, the 100 gig, 400 gig, 800 gig terabit per second -- sorry, the Ethernet technology. So again, here, we need these multiwavelength laser sources and that's not presently integratable in our technology, hence the need for hybrid integration. In the next slide, you will see another example of a technology platform is kind of, let's say, an offspring of our communication platform that is more tailored towards LiDAR applications. So here we integrate high-quality silicon nitride waveguides. And as you can see here, if you click once more here, you will see that we have also here some reference test chips aimed at beam steering. So it is a chip that's operating at 50-50 nanometer wavelength and it's really kind of a reference design to integrate the beam steering element at the chip scale. So you no longer need these bulky MEMS type systems or other type of beam steering elements that typically adds to the size of a LiDAR system. Of course, this is not a full system just yet. It's just a key building block that we have demonstrated and that's also attracting quite some interest from prospective customers. Again here, we have the need for an integrated 50-50 light source to complement the toolbox. Next slide, please. Yes. And here, you can see a final example, which is more in the life science spaces. So here in the life science area, you typically work with wavelengths that are a bit shorter, so between 450 and 900 nanometer. So also for this application space, we have developed a silicon nitride photonics technology platform, which, in fact, can be directly deposited on top of a standard CMOS chip. So you can also see, if you click once more, that is kind of a waveguide layer that can be, for instance, deposit on top of the back end of line of a CMOS technology and that can then interact with silicon pixels embedded in that CMOS chip to make a more comprehensive spectrometer, for instance. And this is a technology that finds use cases in biosensing, cytometry, Raman spectroscopy and also microscopy on chip. Next slide, please. So maybe now turning back to the projects we are running together with our partners, Sivers Photonics and AMICRA. So as Andy already alluded to, we have picked this flip-chip -- high precision flip-chip bonding approach as the most mature near-term solution for bringing lights to our silicon photonics platforms. So the key challenge here is indeed to have very good alignment between these 2 parts, so we can have a very efficient light transfer from the 3-5 components, indium-phosphide component in this case, to our silicon waveguide. So typically, in order to kind of have efficiencies well below, let's say, 50% or even 80%, we need to reduce that placement error to less than a couple of hundred nanometers, typically less than 0.5 micron of precision is required here. So in order to get that and maintaining sufficient throughput of that assembly process, we need to do a careful job at designing that mechanical, optical and electrical interface. So that's why the engineers at Sivers site need to work tightly together with engineers at Imec sites to codesign that interface, and that's exactly what we have been doing for past 2 years. And of course, we also need to complement that then with a tool that is capable of getting these high precision placements done at a high throughput and that's where the collaboration with ASM AMICRA came in. Next slide, please. So exactly -- so this is the partnership we built together. The [indiscernible] to have some kind of a library -- set of library elements developed prequalified, let's say, reference designs that the prospective customers can use and they can also fast track their product development, so they don't need to start from scratch. That's exactly what we have been working towards. And in the next slide, you will see some results that we also published, I believe, a couple of months ago. So we have started this collaboration with 300-meter test design in our silicon photonics R&D fab, where -- so after codesigning the interface between the Sivers and Imec engineers, we have taped out a test chip here, developed a process so we can etch these cavities in our silicon photonics wafers. As you can see there on the top right picture, that's an image showing the cavity etched in the silicon photonics wafer where one of these indium-phosphide [indiscernible] from Sivers has been placed using the AMICRA tool. So we can see here that this a process that we do at die-to-wafer level, so we can really have high throughputs. It's not a die-to-die process. It's really using wafer scale silicon photonics, so we can use this for the more demanding high-volume applications. The next slide will show you, in fact, the alignment precision that we have demonstrated so far. So we are now re-hitting our targets in having the X, Y alignments precision well below half a micron post-bonding alignment accuracy. It's also important that effect -- you can develop this process, but you also need to codevelop the testing processes. Typically testing of photonic devices can take quite some time. In Imec, we have quite some experience in testing, wafer scale testing of our silicon photonics technology. And so we're now also expanding that -- the know-how to doing the wafer scale test of the combined 3-5 silicon photonics technology. And you can see there on the right of photograph of one of these assemblies placed on our wafer-scale prober, where we can now really do LIV curve, so checking how much light is really coupled into the silicon photonics circuit. And we have already achieved more than 10 milliwatts of optical power coupled into this -- on chip waveguides, and we're working further optimizing that coupling efficiency as we speak. Next slide, please. So that brings me to the -- my takeaway. So I hope I convinced you that silicon photonics is really seeing this tremendous growth. We really see this hockey stick trends where we have a variety of applications driving growth now from transceivers, optical interconnects to various applications to sensing domain. We have been able to put together, I would say, state-of-the-art silicon photonics platform to serve many of these applications, but this light source is really missing from our toolbox. And that's exactly what we're codeveloping with Sivers and AMICRA to be able to offer to our joint partners a more complete and comprehensive solution, including the light source and all the other active components. And with that, I thank you for your attention. I'm open for any questions.

Thank you, Joris, and thank you, Andy, for that. Now we've really gone into silicon photonics in depth. And it's really good to have done that, so people really can understand what we're doing actually. So now we're going into the Q&A part, and I'm getting questions here. And we also can take questions in the audience. So I don't know, can everybody hear us now in the -- Atul, do you hear me?

Yes, I do.

Very good. So I'm going to start with a question to you, Atul. How many of your products have Sivers inside today?

We'll be shipping our first Sivers product next quarter, and it is in the POCs with a few customers we have already shared with them in Q3.

Yes. So I think the question was about that you have it in the base station and in the home CPE unit?

Yes. I think as far as I know, both will have the Sivers products. And 5G and R, I think the most important thing is when you look at the 5G, 5G is the first time, at least for us, we are bringing fixed 5G product. Cambium has shipped historically over last 10 years, about 10 million radios worldwide to about 150 countries. Not that all of that will have Sivers, not at all, because they have -- there's WiFi, radios and all sorts of stuff. But that gives you a flavor of the point-to-multipoint, particularly at the edge, there's lot more proliferation, which happens there.

Thank you, Atul. Now questions to, I would say, Andy as well as Joris here. How big deal is the achievement Sivers, Imec, AMICRA and how successful do you think it will be? Maybe Joris can start from your perspective?

Yes, I didn't quite capture the first sentence.

Yes. So how big deal is the sort of achievements that you and Sivers and AMICRA have done together to sort of get to the point you are now? And how do you see this sort of develop in customers and products in the future?

Yes, that's a good question. So we, for sure, have already quite some interest in this project. We -- honestly, we have a bit of a delay due to the corona crisis, so that produced some delay on our side. But we have already customers that are actually working with us and with Sivers on building first prototypes. So clearly, there is strong demand for the capability we're putting together here. So maybe I'll leave at that and leave it to Andy to add more to that.

Yes, I'll definitely that say the offering that we're giving today to customers is really attractive because we take away a lot of the sort of complex interface issues that the customer has to manage today themselves, which can be very complex. So the fact that we have Sivers engineers working directly with Imec engineers takes away all this pain for the end customers. So we think this is a really attractive sort of product offering for the customers. And we are seeing a lot of commercial interest in this partnership.

Thank you. So while we are at you, Andy, here, another question. What's the difference between the sort of 3-, 5-compound semiconductor that Sivers Photonics use versus sort of TCMC (sic) [ TSMC ] and Samsung manufacturer's ordinary CMOS?

Well, they're processing pure silicon wafers, we are processing mainly indium-phosphide wafers, so different substrates. So our substrates are much more specific for photonic applications, silicon [indiscernible] for making semiconductor and electronic circuits. And our compound semiconductors are direct [indiscernible] so you can [indiscernible] from our compound semiconductors, that's why we use it.

And maybe I'll answer that. The question regarding Samsung and TSMC, the silicon industry is [indiscernible] to 5G, of course, is part of that. It's about 30 years ahead in terms of that kind of technology. However, as Joris and Andy had pointed out, the absolute amazing benefits you get of silicon are starting to kind of reach [ a token point ] where silicon photonics will then come in and take that to the next level. So that's why we think the silicon photonics technology is going to be so exciting for the future.

Thank you, Billy. So then we have a couple of questions for Atul here. And you have to answer carefully, I assume. But Cambium, do you give any forecast regarding volumes for 28- and 60-gigahertz products?

No, we don't. We have just 3 categories: point-to-multipoint, point-to-point and WiFi or enterprise. And we -- because our customers buy solutions, our customers rarely buy a point product. They buy edge and transport and complete solutions. So we -- those are the 3 categories that only we disclosed revenue there.

Thank you. So -- and a follow-on on that, what type of customers will Cambium's 28-gigahertz products be used for? Any examples?

Yes, absolutely. Typically, where we are seeing traction are in countries where the 28-gigahertz band is available to service -- especially the midsized service providers because 5G -- fixed 5G is just arriving in a lot of ways because China has just got -- mature chips are arriving. So I think you will see -- you'll see a broad adoption in the 28 gigahertz, 5G countries. We are seeing traction in Middle -- and by the way, every quarter as we release the product, we'll give -- in our earnings release, we will give more commentary. That's our tradition. As we did 60 gigahertz 4 or 5 quarters back, and every quarter, we give more commentary. So you'll hear more from us. And I think EMEA; then probably North America; then CALA, so it was just Caribbean and Latin America; then Asia. In that sequence is what I think will -- the proliferation will happen for us. But EMEA seems to be leading right now for us.

Okay. Thank you about that. Another question then that I also want to understand, why did you choose -- why did you not choose Sivers for your 60-gigahertz product?

I think 60 -- that's a good question and a very clear answer, in 60 gigahertz, we collaborate -- and this is public information. We collaborated very closely with Facebook. Facebook did the [ paragraph ] algorithm, which is a meshing algorithm. And we also collaborated very closely with Qualcomm. And that -- those decisions were made 3, 4 years back and years off, and 60 gigahertz has been around. We did not do the Phase 1 60 gigahertz, which was based on 802.11ad standard because we felt that, that technology needed more maturity at that time. Then came the new standard, 802.11ay. And we adopted or we jumped on that project when ay came, and we felt that there was silicon vendors providing mission-critical solutions. And with Facebook, we felt the meshing, because 60 gigahertz is a small distance and 1 to 2 kilometers. So meshing was very important to bridge the distance. And for that reason, that alliance made more sense to us.

Thank you, Atul. So another follow-up on that, what is the typical range that you -- and throughput you will be handling in the 28 gigahertz, sort of between the BTS and CPE?

The range in our -- a lot of it will depend upon probably different geographies and terrains and weather and all that. Our thinking is it will give us 4- to 7-kilometer type of a range. And both 60 and 28 will be multi-gigabit architectures. That means you can get multi-gigabits out. And as we're doing the proof of concepts that will give us even a better idea in real conditions what we get. But both are multi-gigabit, wireless fabric architectures, both can replace fiber or supplement fiber. And that's why we say, if it can be wireless, it will be wireless. And that's why we say multi-gigabit wireless is the new fiber.

Thank you. Here is, I think, a tricky question. I mean what -- a question to Cambium as well. Please elaborate on pros and cons about sort of owning a larger part of your component supplier ecosystem rather than, for example, Sivers' type of technology, then secure it both in R&D collaborations instead. What's your view on that?

Sure. First of all, we focus on what we do well. We focus -- and that has been our philosophy that wherever there is merchant silicon, use merchant silicon because that gives us economics. So -- and we believe that Sivers is best-in-class in what they do. And so are some other chip manufacturers in what they do. We focus on selecting the best-in-class partner, and we focus on system design. How do we deal with noise, how we do intelligent antennas, how we do beamforming, how we deal with noise, that's -- the entire thing is system design. That's what we do well. So our philosophy is just I -- we believe, like Sivers, they are best in class in what we do. There's no reason. And then when it comes to R&D collaboration, there are areas where we create value in discussions because we are system designers, so we give feedback. But it's not so integral that, that will be an IPR. It's more system. Our IPR is how we put together the entire system and solution, how we manage from the cloud, how we scale that from the tower, how massive number of new users in the city or urban area, how they scale, how a network operator can design a profitable business based on our solutions. Those are our experts. How do we support 24/7 across the globe? Those are the things we focus on.

Thank you, Atul. There's a question for me, I assume. How many of your products have Sivers Inside? Maybe that's for you, but that was 2. But in general, we have 26 design wins, but then, of course, there are multiple products, both CPE-based station mesh node, we're inside. So there is probably sort of 40, 50 products or something like that, that's included in Sivers Inside. Let me see here, other questions. The new chipset that you released yesterday, 24.2 (sic) [ 24.25 ], 43.5 gigahertz, will you be able to put them in the same product so you have, example, a CPE that can handle the whole range? Yes, that's the whole idea, then we can actually do 2 module in the same product and use that. Do you think you can present the F100 customer this year? I don't think we can answer that question really. It's actually up to the F100 customers. So we unfortunately have to duck that question. When will new technology by Sivers, Imec, ASM be rolled out? Market consumer products, how big is the TAM on this technology? Billy, do you have any reply on that?

I think we gave some information in terms of the applications will cover many of the verticals that I talked about from optical communications to sensing. So if you were adding on the term for -- based on [indiscernible] for 2024, it's [ several billions easily ] in that market. We can come back with a more detailed question. But it actually affects all of the verticals that we're working on. There's no one area that will not be touched. And I actually believe that those are areas that we've not talked about yet that it will touch.

Yes. So I can add to that as well. I mean what we've said officially so far is that we assume to have first prototype customers next quarter and then volume in about a year after that.

Yes. So -- I mean we are -- the work that we're doing right now, [indiscernible] we're going to making announcements in the coming weeks, if not next week, and/or in collaborations that we're doing. So that's one of the main that [ we'll ] start. So the activity with Imec, we expect to lead to multiple different business opportunities.

Yes. And that's why we'll have sort of invested a lot in the indium-phosphide 100 platform, which is now sort of together with the partners coming to life this way. And that's a way we always like to work in Sivers, that it's through partnerships we can build things. We cannot sort of be alone in these kind of things.

Exactly.

Yes. And then also a question to both me and Atul, I assume. Cambium or Sivers are part of Open RAN alliance. Is this correct? And what are the reasons why? Also, your view on the Open RAN collaboration and standardization for fixed wireless access development would be interesting to learn more about. So yes, one of the reasons for Sivers' part at least is that we haven't been there. We have, so far, not sort of had the low IF type of integration. We have Zero-IF integration to our stuff. And ORAN products are often sort of things that will be sent through a low IF signal. So in that sense, we have not been sort of part of that because our product has not fitted into an ORAN application. However, with our new product, we can actually fit into ORAN application. So that's basically a technical reason for not joining that from our side. Atul, do you have anything you want to add on ORAN?

No, we are not members. We are watching it. But so far, we don't see as much of a need. And we are a medium-sized company. We have to focus. We have to use every resource preciously. So I think when we commit to these types of forums, we commit long term. And so, so far, we have not seen a major need to be there, but we are watching it closely.

Yes. Maybe a question you can't answer. Will Cambium test new -- Sivers' new 5G-NR chips launched yesterday?

No, I don't think so. We have not tested that. But we are very focused on, right now, getting our product out, working with customers with what we have and learning from that. And let me just give like 1 minute quick life cycle duration. Typically, a customer will use a proof-of-concept for 4 to 5 months. Sometimes 6 months because they'll deploy, they will test configurability, ease of deployment, diagnostics because they're going to offer a service, and the service has to be high quality. So it's a -- rather than giving them too many products, too many technologies, it's important to give them a solution, let them work, let them feel comfortable. Then they deploy that in a small setting. Then they deploy in a city. Then they go multiple city. It's a multiyear program with a customer. And so that's why we call it land and expand. So our focus is never to really put too many products out there. Focus is whatever we give that's complete and that's a good solution. And starting Q4, that's what we will do, focus on what we have and then keep expanding from there. We learn a lot from these customers. And some of that feedback, we -- relevant feedback, we pass it to Sivers as they do their road maps of the future.

Yes. Thank you, Atul. So how big is the broadband fixed wireless access market share do you think Sivers can reach? So -- I mean we don't do forecast in general. But as I mentioned, I mean, the overall TAM is about $7 billion. So it's a huge market to address. And so far, we are just in the start of that market. So we expect to be an important player. And together with Cambium and others, we're going to capture some interesting market share, of course. Another question here. Will 6G standard be -- standardized as 3GPP Release 18 and onwards, similar to 4G and 5G? Yes. For sure, 6G will get there. And I think there's also be sort of what we will call the next generation 802.11ay things where you really get to 20, 30 gigabits of things. It's an even lower latencies, which will match 6G in that sense from the WiFi standpoint. There was a lot there. When do we expect more information on the F100? Yes, that we already have answered. So now we have no more questions coming in here. Maybe we can, too, take some questions in the audience here. I see Klas from Nordea here wants to...

So -- I mean, Anders, you spoke -- you hear me?

It's a bit low but okay.

Okay. So -- I mean, Anders, you spoke a bit about your TAM in the -- earlier in the presentation. And I think you mentioned that you can [ turn next that ] basically by moving into new applications and so forth. So -- I mean could you give us any examples of what types of applications that is? And what do you kind of need to do differently to access those markets essentially? And is it only on an inorganic side you're willing to do this? Or do you also see some organic kind of investment opportunities into that?

I definitely think that it's an M&A type of thing that you want to -- if you want to move into that quickly. For example, we have not gone into the handset side of things here because that's the huge number of chipsets in handsets, of course. So the right acquisition where you can get into that is, of course, giving you a much bigger TAM immediately. That's part of the TAM mix. Another thing is this kind of repeaters or SATCOM that is coming in, which is a very close area. We could actually organic grow into that by sort of changing our chipsets a bit. It's a long investment. It takes a long time. You can accelerate that via sort of good M&A activities as well. So there is many options in the different types of applications here.

Yes. And I mean in what type of time frame are we talking here? Is this a next year thing? Or is it 5 years from now? Or what kind of time frame...

If you talk organic, of course, that takes some time. It's a couple of years. When it comes to consolidation or acquisitions, that can happen anytime.

Yes. And I mean looking at a lot of these new kind of applications and so forth, would you say [ they are at ] the same kind of level of maturity as your current ones? Or would you kind of be integrating into a more mature market already where the sales opportunities further...

I think that the millimeter wave market in general or for all of those things that's connected to 5G, like 5G repeaters, handsets and all that, that is sort of in the same kind of level. If you look at SATCOM, that's maybe 1 to 2 years later. But also even further out, all the LEO SATCOM, [ SATs ] and all of those things come out, that's a bit behind. Then you have the defense industry that's often a bit more conservative. It could take a bit longer for 5G in there. But I mean it's within 3 to 5 years as well in that area.

Yes. And Anders, a bit of a boring question maybe. But I mean you're entering a rollout phase now and obviously, there's quite long lead times on chipsets and so forth. And I mean you'll need some high -- quite a high inventory to deliver, I'm guessing, on some of these projects. So could you maybe elaborate a bit about what are your kind of current capital -- working capital needs during the coming year or 2?

Yes. We have actually not sort of gone out with that, sort of the exact working capital needs. But what we have said and what we do to secure chipsets for customers like Cambium and other, [ is that ] we are investing, of course, in stock to make sure that we can buy and have stock in -- in-house to secure the volumes for the future, to not end up in a sort of lack of components. So it is, of course, working capital that we are putting into that as we speak.

Yes. And I mean could you maybe help us understand how many quarters do you have of kind of covered capacity in your mind?

I mean we need to work 3 to 4 quarters out because it's a long lead time item. So it could be everything from 4 to 6 months to sort of get more chips. So we have to work in that kind of way.

Okay.

Thank you. So while we're sending the mic around here, if anyone puts up the hand in the air. We have a question to Atul as well. How does Cambium see on the chip shortage under this -- during this year and generally in the market?

I think as we -- I'm going to stay with this -- since we are in the quiet period, I'm going to stay with the things that I mentioned in the -- our earnings call. In general, the supply chain -- and it's not just last few months, I think it was building over the last few quarters, supply chain, chip constraints has been tight. There is no question across the industry. And it's not even just that. I think if you look at the -- everybody read the news about container ships in California coast waiting. So there's a downstream impact as well in just the demand has increased post-COVID, and the demand has increased from automotive. Everything is going digital. Consumer electronics, people are buying more. So I think this is something which definitely is in -- and different chips are impacted to different degrees. And then there is technological transitions going on as well in the WiFi world, from WiFi 5 to WiFi 6. So again, more demand there. So I think this is something all of us have to just work through in next few quarters. And -- but everyone is putting more capacity. Everybody's working on it. Everyone knows that there is chip shortages and impacting many segments right now.

Thank you, Atul. A question here regarding the Ampleon work stream and how it works. And of course, Ampleon and us have had 2 projects. One is the chip that actually is in Cambium's products today. So that's been a joint venture to bring that chip out. The other one is the -- which I think the question is about is the beamformers that we released about a year ago, and they are now out in sort of test equipment and people are trying them. We had, had a few press releases about -- in Taiwan and India and so forth for the [ testings ]. And they're still in testing, and they are not sort of in any design wins as of yet. Yes, any questions in the room here? Yes, we have [ Fredrik ] over there. So while the mic is going up to [ Fredrik ] first from Handelsbanken.

Thank you for the presentations.

Thank you, [ Fredrik ].

On the Photonics side, maybe Anders or William, feel free. If and when you receive and get these volumes started, what type of production capacity do you have in place? And what type of investments do you see in front of you during what time span that sort of fits into the schedule you have going forward?

So that is, of course, a very sort of secret question in some ways. But as we've said before, I mean, to start off the duplication of lines and all of those things. I think we've talked about officially, so far, $30 million or something in that line. Then of course, if we're going to sort of really build out and invest in a big fab, we're talking even more than that. I don't know if you want to elaborate more on that, Billy, without saying too much.

Yes, yes, sort of a leading question. But basically, we have still some significant capacity in our current facility. And so we -- based on the way the transaction will happen is for the next year or so, and we will still be transitioning in this facility, expanding here for the pilot. But as Anders says, that's your time period for that -- from kind of start to beginning is about a year. And so basically, the first pilot line would come up in a year from a 0 point, I would say. And then we would do that in a modular fashion so that we can then -- we have already worked out the equipment since, [ the floor plans ] and average everything we need to do with that is a fairly simple case of [indiscernible] facility. And -- but I can't really get into any more detail on that. It's fair to say that we've shared all of those details with the Fortune 100 companies from [indiscernible]. So we are actively engaged in those conversations with them about what that expansion will look like. And that's the kind of level of discussion we're out right now.

Thank you, Billy. Yes, [ Fredrik ]?

And maybe a question to Atul. You had a few slides on customer cases that I'm not sure that I saw it straight on, but you referred to a case where you're replacing a traditional microwave backhaul network. So is it fair to say that your product where -- that you're replacing it with is a microwave backhaul? So you're sort of competing with Ericsson and Huawei products? Or are you sort of disrupting that backhaul part of the network?

A very good question. No, we are not replacing microwave in general. I think microwave does what they do very well on distances. What's happening is that if you are in an urban environment and you're putting video surveillance, for example, and you have to go around the buildings, you need more meshing. So those situations, the 60 gigahertz become very cost effective. So that's a situation where even if somebody had microwave -- since microwave is a licensed frequency, it's a different architecture, where a 60 gigahertz is unlicensed or likely licensed depending on the country, you can go around the building. So there are applications where 60 gigahertz -- whether it's a small cell architecture where you have to backhaul with high speed and 60 gigahertz can give you that small distance and meshing capability. So I think in general, 60 is unleashing new applications, not necessarily replacing a legacy stuff. The new applications is high-speed broadband. So if you are in a high-density urban environment and you need to bring broadband, you don't need permits, you don't need to dig streets, you can just bring quickly. So I think 60 is a new building material. I always tell people when steel came at turn of the century, 100 years back or so, you couldn't predict where all steel will go. It's a new building material. It has gone into health care, the surgical steel, they are all sorts of stuff, that's the example. I think 60 gigahertz and millimeter wave in general is kind of extension of that LAN in multi-gigabit onto the distance. And that's what we say when we say wireless fabric in the last -- on the edge, multi-gigabit wireless fabric on the edge. So no, we are not replaced -- I think there might be applications where some microwave may get replaced, but that's not the main thrust.

Thank you, [ Fredrik ]. We have [ Victor ] from...

It's yet to be announced, a fund manager. I have a question mostly, I think, for Cambium and Sivers. There's a lot of buzz around private wireless networks now. I was just curious to get your take on what kind of traction you see in your business in the private wireless networks or any other thoughts on this topic?

So Atul, do you want to start?

Yes, I can. I can. Thank you. Private wireless networks are -- that's what Cambium does. Actually, most of our networks are private wireless networks. A substantial part is provided by the small- and medium-sized [ risk ] wireless Internet service providers across the globe. But more importantly, our networks also go into enterprises. As I said, oil and gas, mining, energy grid, water and waste management, smart cities, and it keeps expanding. Post-COVID, everyone is focused on scalability because there are lots of performance and media-rich applications, which are driving it. And also security. The -- so I think, in general, there is more proliferation of that, controlling the network, controlling quality of experience. So we see definite expansion in private networks. And what 5G is doing? 5G is legitimizing it. 5G is saying that both license frequencies, unlicensed or likely licensed, they'll coexist. Not all applications need that mission-critical license, expensive on a 5G. Many applications will need that. Quite a few applications will be mission-critical enough with fixed 5G, and that's why we see proliferation of private 5G networks across enterprises, small cities, municipalities, localities, you'll see all of that. And each network will optimize the application. You'll see more marriage of low latency, high throughput, uptime, downtime, everything [ that has a ] little video surveillance use a lot more uptime -- uplink speed. That customization happens through private 5G network, a lot better.

Thank you, Atul, and I can add to that as well. I mean if you talk about enterprise as sort of industry 4.0 type of applications, where both Nokia and Ericsson is talking now about licensed products and so forth, I think that unlicensed products in general in that application is much easier to use rather than sort of expensive spectrum as well as looking at track-to-train application, that is actually a private network in a sense, and they don't really need licensed. They need a dedicated track-to-train system that they can run license free. So I think there's a lot of things in licensed -- unlicensed spectrum that could be used for a lot of this. That -- Nokia and Ericsson is trying to hype the license bands as well a bit more.

Okay. And take one more question also for Anders and Billy then. On the optical sensors part, I noticed Billy talked a lot about AR and VR. Is this the most exciting opportunity you think? Or is there any other super exciting opportunity that you want to highlight in this space?

If I can answer first, and I think Billy agree, I mean it is all of those spaces and what Imec's talked about today, optical interconnect, LiDAR, augmented reality, all the different pieces here, where you actually can use silicon photonics. And with this, as Andy said, small tweaking in between, and it's actually easy to then use it without sort of reinventing the wheel the whole time. So I think it's a lot of the spectrum there, which makes silicon photonics so hot right now. So if you want to add something, Billy?

Yes, it's a really good question. Actually, although I think augmented reality, mesh reality and [indiscernible] to be utilized in an industry, that's going to be used in space, it's going to be used in so many support systems. However, I think sensors that will be applied in kind of bio functions such as wearables, glucose monitoring, toxicity monitoring, basically having health systems attached to things when certainly on person, that's extremely exciting area. And I think we are doing a huge amount of work in both of those kind of -- sort of verticals at the same time on both of those areas. And I would say probably the kind of buying sensing stuff that's going to change the health industry, it's probably the most significant, and that will be a game changer for everyone.

Thank you, Billy. So we are actually at 5:00 now. We can have one more question here, and I'll take one question from the online audience. Anders, how is the remaining recruitments in the U.S. office going? And as you know, we have just recruited a VP Business Development for Photonics. We are actually now in the end of the recruitment for the wireless piece of the business. So we have a couple of candidates we're looking at, and hopefully, we'll know within Q4 who's going to enter the role there. So this is a very exciting time on that. And then after that, when we now recruited these 2, we look at the 2 other headcounts. So it will be sort of presales engineers that they will recruit. So that's the situation there. So okay. So that was all for today. Thank you so much for spending so much time with us. For you guys who are here, we will mingle a bit outside here and drink water or possibly wine as well, if you like, and you can ask more questions. And for everyone online, thank you so much for listening. I saw earlier, there were several hundred people online actually listening. So thank you for joining us. And thank you to Atul and Joris at Imec for joining us at this. It's been a pleasure having you.

Thank you very much. Have a good day.

Thank you. Bye-bye.

Thank you. Thanks, everyone. Thank you.