Transcript: POET Enters $200M Market for Long Reach Optical Engines with Advanced Customer Discussions

George Tsiolis, 2:14: Gentleman, welcome back.

Thomas Mika, Vivek Rajgarhia, 2:17: Thank you.

[2:19, GT] Vivek, there’s a lot of tech in this press release, so we’re going to focus more on what that actually means from a tech point of view to the layman, and what it means to the business. First of all, the quote here is you’ve completed the design of a 100G LR4 optical engine with a reach of 10km.

[2:41, VR] Correct.

[2:42, GT] What does that mean? Why is that important to any of us?

[2:45, VR] So, George, let me step back a bit to describe where 100G LR4 modules are used. So it’s actually one of the first standards for 100G transceivers for connecting what we call the client side. As the fiber optics’ demand and requirements are increasing—the increase of data—they need to be connected, within the cities, from city to city, from country to country, and then all that needs to come closer to us, the consumer. This equipment, which is connecting the cities, the countries, is what we call the long-haul [line side] transmission equipment. There is a side to it, which we call client side, where those connections need to go into data centers, into enterprises, into close to our homes, and so on. Those connections are used using 100G LR4. So as the line side transmission equipment demand is increasing with coherent, and so on with DWDM, we need more and more of these 100G LR4 coming closer to the client.

[4:05, GT]: Naturally. Naturally.

[4:07, VR] That is where it’s used. The main difference… in the past we’ve talked more about inside the data centers, CWDM4… so “C” is coarse, ok? Which in uncooled. The total four channels that are covered by a CWDM is about 16 nanometers—that is the different colours of light that go in that band. In the LR4, all those four channels are covered within 15 nanometers. So much shorter band, more precise, more cooled, with higher performance so you can even go 10km or longer distances. And this is a de facto standard which, as I mentioned, was early on for 100G, and which will remain for a long time because of interoperability, and because of the performance levels that are required to connect the transmission equipment to the data center. You need good performance and precision at that side.

[5:12, GT] That makes sense, because there’s just more demand, right, at the end of the day? More surfing, more streaming, more everything at the end of the day?

[5:18, VR] Exactly.

[5:20, GT] So let’s… you’ve laid out, guys, three advantages: one being that a 100G LR4 transceiver sells for 2-3x the price of a CWDM4. So that’s one advantage. Does this represent an opportunity for POET? Or is it just the cost is 2-3x higher, so you’re selling them 2-3x higher, and there’s not much of an advantage?

[5:45, VR] No, actually, it’s a very good sweet spot I would say for POET. CWDM4, as I mentioned, it is very high volume. It’s used inside the data center of the hyper-scale service providers. The LR4 connects those to the transmission equipment. The prices are higher. Volumes are slightly lower than CWDM4, but because the performance levels, the power consumption levels, are much more stringent, we provide a good… it’s a sweet spot for us. Also, the costs have kind of hit the bottom where the cost of the modules by these providers cannot go down further… it’s reached a certain level. But the prices need to continue coming down and the margins are being cut into our module customers. So we can provide them a solution where they can continue a viable business, keep making money in a stable and growth area, and provide the functionality and the performance levels that are needed.

[7:00, VR] One thing, I’ve been a part of making these LR4 engines in a more traditional manner. They are very complicated, very complicated because there are certain requirements needed. It’s a complex module: the alignment needed, the lens, the precision of alignment, how much power you can get, how you manage the power consumption… it is very complicated. There are a few number of providers for these engines today, even in the traditional method. What we are doing—what POET is doing—again coming to our optical interposer, where monolithically the waveguides, the mux and demux, are all on the interposer, they’re monolithically provided in the CMOS fab. Then we flip-chip passively align lasers to this, which removes all the complications of doing the alignment using lenses and epoxy and such. All those mechanical alignment features are already on our interposer, so you flip-chip attach, passively align these DML lasers to the MUXs, and get the light out. This really provides a way to reduce the cost and pass part of the cost savings on to our customers so they can make money in this mature market.

[8:22, GT] And you state savings to customers in the range of 25%?

[8:26, VR] Yes.

[8:27, GT] Look for most products if you can save 25% that’s significant, so I’m assuming that’s the same for your customers. Is that a massive competitive edge? Because it sounds like there’s a lot of competition in the space… or if not a lot, very intense competition. So how big is that savings to customers, when you’re talking to them?

[8:47, VR] So, the way I would describe the competition is that it’s not like CWDM4, where there are many more suppliers. There are limited suppliers in the LR4 because of the complexity, as I mentioned. But I would refer to them as the incumbent suppliers, but those incumbent suppliers are hitting a bottom—or have hit the bottom—where they can’t reduce their cost and their margins are diminishing. This is where we step in and provide them a solution where they can continue to make margin. So making 25% more margin, versus not making margin and trying to stay in the business, you can imagine this is big…

[9:32, GT] This is massive. All of us at home as investors, there’s no way 99.9% of us understand what the competitive landscape and the factors are but, Vivek, it sounds like what you’re telling us is that’s a massive competitive advantage when you’re walking into George-Com and saying, you should be using POET versus ABC-Technologies.

[9:54, VR] Yes, exactly. And it’s not a cluttered market, so we have very targeted customers. Some of those customers… so, let me come back to the opportunity. Even though the number of suppliers is limited, it’s still a billion-dollar market. It’s not a small market.

[10:15, GT] And growing? Vivek, I assume it’s going to be growing?

[10:18, VR] It’s not growing like a hockey stick, it’s stable, it’s growth… that’s why I call it a sweet spot for us because it’s a stable market which will continue for several years to come. As the growth for data continues, this will be needed more and more and it provides a very good opportunity for POET.

[10:41, GT] What’s an example of a kind of customer that would be buying this… and I know you’re not going to name a customer, just give a name that people would understand, that this would be potentially someone who would need these products.

[10:54, VR] Yeah, I’ll describe the value-chain a little bit. So we are providing optical engines, which will be manufactured at our joint venture, Super Photonics, and then our customers are the module providers: companies that are making/integrating modules, QSFP28 100G modules, and then they are supplying to telecom equipment providers. So these are companies—again, just as generally naming companies—companies like Nokia, Ericsson and, you know, ADVA, those type of telecom equipment providers buy those modules and then supply it more in the telecom networks, which are connecting to the data centers. Also data center companies’ providers would buy to connect to the telecom equipment, but a lot of the demand goes to what we call the telecom equipment and network providers, versus the hyper-scale data centers to use inside their data center. So that’s a slight differentiation here.

[12:10, TM] Let me just add to what Vivek has said. Remember, a key to our strategy is that we’re not trying to replace, or compete directly with, a module manufacturer. What we’re doing is providing components within the module. And the components that are being sourced currently, are being sourced from hundreds of suppliers. So we’re not competing with a monolith company, that is more powerful than we are. We’re going to the module makers saying take those hundreds of subassemblies, or tens of subassemblies, that they have in this module out, and substitute for it our optical engine at a much lower cost, and likely a much lower assembly cost for them. And, in doing so, we can reduce you BOM [bill of materials] cost and your assembly cost and provide you with… re-establish the gross margin that you’ve given to the telecom equipment suppliers over the past few years.
GT, 13:25: So, guys, I would assume… and I’m going to go into something more specific, but given what you just said Tom, which was great, I would assume that you guys already would’ve had some back-of-the-napkin conversations with potential customers, with the module makers, what kind of feedback have you gotten from your target market?

[13:48, VR] So, actually in this case we didn’t even have to proactively go after customers, the customers came to us. There are customers we mentioned in our press release that we’ve engaged with… there are a few customers. We are working very closely with the customers to find the certain level of customization needed, which we can do with our interposer. We are doing that. And we’re in active discussions with customers, so we will be working very closely with the customers as we bring out our samples. There are customers who are already lined up and are waiting for our samples to evaluate. So I think it’s a very good position for us.

[14:37, GT] Hey, even to the layman, we understand that. If you’ve already got customers already lined up and interested, you’re way further ahead than George-Com who develops a product and then I’ve got to go give out samples and hope somebody likes it. That’s fantastic. So let’s talk about timing. Obviously the first step is you… it sounds like you’ve already gotten samples out to customers [Tom Mika shakes his head, no], and you’ll be continuing to put samples in the hands of customers. How long does your profiled customer take to evaluate your LR4? Is it six months of time? Because, at home, none of us know that. What does that next stage look like?

[15:23, VR] Again, a good question George. The sales cycle, from sample to revenue—right?—that’s really what you’re asking. So there is a process. This is an established market, and as Tom mentioned earlier, we’re not saying instead of walking you have to fly here. We are not changing the approach to how it’s being used, we’re just changing how it will be integrated into our customers’ (products) in order to save costs for them and increase value to them, right? So, this standard product… again the overall cycle time… our customers also have customers already that they have been supplying to, you know, for different configurations. So generally, 6 to 9 months from sample to revenue is a rule of thumb that we have used in the industry. The specs and all are well defined. So its not like we are going to be trying out... the customer is saying “no, the power needs to be different or something.” So those unknowns are not there.

[16:33, GT] You’re not at the trial and error stage anymore.

[16:35, VR] Customers still need to validate their modules. We have a good team to work closely with customers. We have demonstrated this in previous engagements. You know that large European telecom equipment provider, that we had announced some time back, that we are getting out, and getting into their modules. So, we have a good setup and experience of working closely with our customers.

[17:06, GT] Thomas you wanted to chime in there with something.

[17:08, TM] Yeah, I just needed to correct something that you said earlier that customers have samples of our devices, they actually don’t. What they have are our designs. We don’t expect to have samples until Q3. One of the things that we pointed out in the press release… There are actually two things. One is that we’ve had these wafers that this device is being produced from in fabrication since December. This shows you how stretched out the supply chain has become, where we don’t actually have those wafers yet. We expect to have them soon, but we don’t have them yet.

[17:53, GT] Is that a potential weakness, weak point, for this operation, Tom?

[17:58: TM] Well its something that the entire industry is experiencing, right? The semiconductor industry is experiencing. So all of these foundries are delayed including ours. But the countervailing thing is that the lasers that we are expecting to use are being provided by SAIC, our joint venture partner. And those lasers have already been qualified and our design is in qualification. So on the one hand the interposer wafers are somewhat delayed, but the lasers which frankly we’ve been delayed on longer than we’ve expected, for several months, for other applications, are going to be available sooner to us. So that will help shorten the qualification time. But nevertheless, we look at the plan and see that we won’t really have samples that we can turn over to customers until about Q3.

[19:03, GT] Okay going by what Vivek said earlier, Q3 you start getting samples into the hands of customers. And if it takes 6 to 9 months, that’s 2 to 3 quarters. We’re talking about Q1/Q2 of 2022 you should expect to start seeing… And this isn’t binary, right Vivek? It’s not as if either they like it or they don’t. You definitely expect to be gaining some market share here.

[19:22, VR] Yes, yes. There is some feedback from customers after we sample. I expect them to be tweaked when we go from alpha to beta. Again, these are very well-established standards for these products. So, the unknowns are not there. So middle of next year to get revenue, production revenue, is very realistic for us, for this.

[19:54 GT] What do you guys think? If it’s a billion-dollar market, what’s a realistic percentage of that that you guys would be happy to.. obviously, you’d be happy to be 100%. But what’s a realistic number where you’d say… By the way to be clear for everyone at home, you’re not providing kind of projections. But what’s a realistic expectation of market penetration.

[20:15, VR] So let me clarify. The billion-dollar is for the transceiver, our customers, you know. serviceable market. We have a portion of that content that we provide. In the past we’ve said anywhere between 30-50%. In this case let’s call it 40%. So a $400 million market, addressable market, for us. There are certain customers who are vertically integrated who do everything themselves. So those are not going to be necessarily our customers we are going to enter with. However, I do have the expectation over time, as their margins diminish, that maybe they come to us. So, they exit the market. So a portion of that customer base that we… that is wanting to source from us, that we target, Let’s call it 50% of that. $200 million. That’s really the size of the opportunity that we believe realistically in a few years that we can get to.

[21:23 GT] Do you think you can capture all of that $200 million in a few years, or would you be happy if you captured 30%, 40%, 50%?

[21:31, VR] So, of course we would want to capture all, but realistically there are not too many customers, as I mentioned. Its not a cluttered business that we have to service 20 customers. There’s a handful of customers that are meaningful. For that about 50% is a realistic goal in my mind.

[21:57 GT] On that note gents, you have an important industry conference coming up. Which I assume you are going to be utilizing to showcase this amongst other things. What are you guys planning to exhibit at that event?

[22:12, VR] We are working,…actually really positioning. Tom and I have been talking almost daily on how to do it. It will be virtual which is [garble]. And it will are setting up our display almost as if we would be in an exhibition room, or in an exhibit in a conference. We are setting up live demos to show customers. So we will be able to stream for customer meetings. We will have a page on the OFC website for customers to access, for people who are walking in, as if they are walking into an exhibit. They can approach us if they want to schedule meetings. So I am actually quite excited about… I’ve been attending OFC for decades and I’m actually quite excited. This will be my first virtual exhibition. But there’s a lot we can do, and we can have the team in one place… we can have people in multiple locations stream at the same time and join meetings as well. So through Zoom and through others. So Tom, you want to say you’re leading this OFC effort.

[29:32, TM] Well it feels like I’m really behind in it because there is a lot to do in a short period of time. But we’re intending to not only have the live demos but also to have some videos available for people who visit the website. If they can’t arrange a live demo via the streaming, over that week, we’ll have a video of the live demo to show with some animations and a professionally done video that will feature the interposer, feature the optical engines, and the products we intend to announce either before or during the convention.

[24:13 GT] That will be on the Poet site Thomas. Will that also be available for shareholders to watch after the conference?

[24:20 TM] Yeah. It will be. We will first have it at the OFC, on their page, and we’ll port that over to the Poet website. I did want to add one thing. In our last conversation, Suresh really recommended that shareholders focus not on revenue for 2021, but on design wins. And so, we expect to have those design wins, which is really an indication that our technology is being taken up and that revenue will follow. And those are the kinds of announcements that we expect to make for the balance of this year.

[25:05, GT] Thanks for stressing that Tom because I think it’s really important and that’s why we put some time into that part too. For everyone to understand. You know, there is a process. You’re not selling widgets on the street corner where you get people handing you cash. There’s a process. But you guys, Vivek, Tom, are really happy with your cadence so far as a company with these products and how you are getting through the process.

[25:28 VR] Yeah, definitely. I feel quite encouraged with the progress, with the opportunities. You know the value we can provide with the joint venture being able to scale with Sanan IC providing devices, at least in the competitive markets of 100/200G. The only, I would say, something that could have gone better, you know, less delays from foundries and all. But as we’ve indicated before, it’s not impacting our business case, business forecast. But you know we want to sample customers sooner rather than later. And also, we want to see the fruits of the efforts that the team has been making.

[26:13 GT] You’re like the latest great horse at the Kentucky Derby. You’re in the stall and you want to get out. And that’s great. I love that confidence and I’m sure that comes with everybody whose watching it. That says a lot when you are anxious to get going because you have a very high expectation of success and that’s always a good sign. Tom, last question to you I would think, because you are the CFO. Outside of products and design and all these things, how is progress going towards the Nasdaq listing?

[26:45, TM] Thanks for asking, George. Actually, its going well. We have a lot of ducks to line up in order to make that an effective launch. Essentially, we’re looking at the Nasdaq listing not just as a listing, but in effect as an IPO in the Unites States. That requires some restructuring of how shares are transferred, who our transfer agent is. We have made application to the Nasdaq. We have received comments back. We’ve got two teams of lawyers involved and providing responses to that. So, it’s going to take some time, and as we’ve said previously, we’re going to do it when we believe it’s the right time for us to do it.

[27:36, GT] Yeah, and I would assume, Tom, that involves some wins and things like that so you can go in with some momentum. Where you know that what your expectations are are actually happening. Right? That would seem logical.

[27:51, TM ] There are a fair number of institutional investors who would like to see some of the risks that still exist to be diminished, right? And design wins really are what’s going to be important to demonstrating that.

[28:10, GT] But what I like about your strategy, gentlemen, is that you’re already getting the process, the progress going towards the Nasdaq listing. You are doing everything concurrently, as opposed to consecutively. So you are not waiting for those risks to be mitigated, design wins and so on, and then starting. It sounds like your doing both at the same time. You’ve got two tracks. You’ve got the business side and you’ve got the capital markets side, going on the same track. Fair to say, gentlemen, both of those, the overall picture, you guys are happy? And should your shareholders be really happy with how progress is going as an overall, one, big entity now? Everything moving forward?

[28:54, VR] Yes, definitely. You mentioned OFC. Actually, I’m in our Allentown location. Tom, Suresh, we’ll all be here doing the streaming. So, our critical mass is increasing. We are, I believe, very well positioned to take… you know, capitalize on our technology that is now coming out in terms of products. So, yes.

[29:22, GT] Gentlemen, thanks so much for today’s interview…