A Christmas Joint Venture -- Flip Chip Video transcript
posted on
Dec 24, 2020 09:44PM
First, get well Sula... Agoracom's soldier/wordsmith.
InThisThinRain and myself collaborated on the latest transcript. Sharing the load means errors may not be mine. ITTR did a great job except he spells color like "colour". :)
The Agoracom youtube video does have closed captioning, but its nice to have the full text. Also note that at one point Suresh said "we are fabless" but closed caption said "we are fabulous". Maybe he did say fabulous.
Rogue
George: Welcome to “Beyond The Press Release”, a production of Agoracom where we talk to small-cap executives right after they put out important news. The company we are talking about today—talking with—Poet Technologies. It trades on the TSX Venture exchange on the stock symbol PTK and, for our friends in the US, under POETF—a great symbol—on the OTCQX, which is the most senior of the OTC markets there. With us today, all three of the gents are back: Suresh Venkatesen, chairman and CEO; Vivek Rajgarhia, president and general manager; and Thomas Mika, executive VP and chief financial officer. For those of you new to the story, because I said this last time, we’ve got two kinds of people that are going to be watching this: 1) long-time POET shareholders—love having you here; and 2) given the fact that it was another big day for POET Technologies—closing up twenty cents… sorry 20%, closing at $0.72—it means we’re going to have a whole bunch of new people watching. So, to understand—because we’re all going through this—to understand POET Technologies you first have to understand “you” (points at the viewer). And what I mean by that is everyday people around the world, you, us, we’re binging Netflix, watching YouTube videos, shopping on Amazon, streaming Spotify, and even Zoom-ing meetings like this. And we do it all demanding the fastest speeds possible; anything less than the speed of light has basically become unacceptable. But that’s not all we demand, we want even more than that! While we’re doing all these things we want artificial intelligence to know our preferences, and recommend us the next movie, the next song, the next video, the next travel destination, or whatever the case may be. And, again, anything less than the speed of light plus the kind of super-computing/cloud-computing power is unacceptable nowadays. All that power comes from servers, data-centres, cloud computing… but what really connects all of them? What transfers the speed… what transfers the data information as fast as possible: the speed behind it? That’s where POET Technologies comes in with their photonics, specifically photonic devices. These photonic devices create, detect, and manipulate light. Laser generated light is critical, fundamental, to sensing, computing, telecom/telecommunications, all of which require the fastest transfer of data possible. What POET has done is develop a unique, disruptive, differentiated new entry into the photonics market: they call that the POET Optical Interposer Platform. And we know that sounds a little bit complicated so that’s were going to talk about partially today. But the press release we’re talking about is, “POET Announces Industry-First Flip-Chip DML Lasers.” That sounds complicated. Guys, let’s talk about it. Welcome to the show.
(3:10) Suresh: Thank you. I’m here with Vivek and Tom as usual. I think we’re going to spend the next few minutes here explaining what we’ve done as a company and what it really means to us as a company and you as a shareholder, either as a long-standing shareholder or potentially new eyes looking at the company as a fantastic new investment opportunity. So before we get started, I kind of want to set some basic definitions out there so we can understand the rest of this presentation. We’ve used many terms in the past, and even in this press release, so I think it’s useful and important to articulate what that really means. You know, the first thing we keep talking about is the concept of a “platform technology.” What does a platform mean? I think I usually go back to an automobile analogy, because that is one industry that has embraced the concept of a platform as a means to quickly put out derivative products, a means to have efficiency, the means to provide scale… and, for example, you could have an engine and you could have multiple components: a six cylinder, a four cylinder, what have you—you’d have multiple power levels. You could have a front-wheel drive, (or) a rear-wheel drive transmission. You could have a chassis that can be put together… the same chassis can be used in a sedan, it can be used in a SUV, and can be used in a sports car in a two-door version. So that’s the definition of a platform: the creation of the chassis, the engine, the transmission, the wrapper, if you will, around the product. And so that means when a customer asks for XYZ you can put these things together faster and more efficiently than you could otherwise. So that concept is what we’re adopting here at POET when we call something a platform. So, the fundamental interposer is a chassis on which multiple components are assembled. And that chassis itself can take various forms, but that is one basic component of the platform. Just like engines power a car, lasers power our interposer, so we would have different types of lasers. What we’ve talked about in this press release, it’s called a directly modulated laser. Or we’d have other lasers. Lasers can have different power levels just like different power levels in the engine of a car. Some people might want something that puts out 20 milliwatts, some people might want something that puts out 50 milliwatts, so this portfolio of specific lasers is important to allow us to customize solutions for customers. Similarly, with transmissions, we have filters—we have two channel filters, four channel filters—just like you would have with a front-wheel drive or a full-wheel drive. So we are creating a menu platform that allows us, over time, to quickly create derivative products that address larger and larger and expanding markets. So that will hopefully give you a good picture of what we mean when we say a platform.
(6:35) Suresh: Now let’s talk about lasers in particular. In this particular press release we talk about a directly modulated laser. So think about a light bulb. Ok? You turn the bulb on and it’s on and you turn it off and it’s off. Everything that we’re doing is digital signals, which means you either want to transmit a one or transmit a zero. When the light is on, you’re transmitting a one. When the light is off, you’re transmitting a zero. So the whole point is can you modulate a particular light signal between ones and zeros as desired by the information you’re trying to transmit? In the case of a directly modulated laser, it’s turning the light switch on and off. So you’re directly turning the light switch on and off in order to create either a one or zero. That is contrasted with what we call a continuous-wave-laser, which is externally modulated, and in that case the light is on all the time and, let’s say, you take a sheet of paper and you either put it in front of the light or away from the light so you’re either blocking the light or transmitting the light. So the light is on all the time and you’re modulating the light externally. That’s called a CW, or continuous-wave-laser, modulated externally. And we’ve got both of those types of lasers as part of our platform. So a directly modulated laser is a laser which is turned off and on as you’re transmitting ones and zeros and the frequency, or the speed at which you can transmit that information, becomes an important metric. So the lasers we’re talking about transmit at 25 gigabits per second. One gigabit per second is one-billion bits of information per second.
(8:26) George: Unbelievable!
(8:28) Suresh: One billion! So our lasers transmit at 25-billion-bits per second. Now as you get to frequencies that are even faster than that, that is 50-billion-bits per second, then it becomes very difficult to directly turn the light switch on and off that fast. It’s much easier to modulate it externally, and that’s where a continuous-wave-laser comes into the picture because those lasers are applicable to higher frequency applications: 50 gigabits per second a beyond. So our portfolio is comprised of directly-modulated-lasers that can go up to 25 billion bits per second, and then continuous-wave lasers that then take it beyond that to the next generation of technology, which is the 50-billion-bits per second, which are used in the 400G data centre applications. So that’s just sort of a broad example or description of what it is we’re going to be talking about in the press release.
(9:28) George: That’s a great example. A great example how you laid that all out.
(9:34) Suresh: And then the final point I wanted to make is, what is the term that we call flip-chip? In the semiconductor industry, over decades, you create a semiconductor chip—and that chip can be a micro-processor, it can be a micro-controller, it can be a memory chip, it can be a laser or a photodiode, it doesn’t matter what chip it is—that chip essentially has to be put into a package before it can be put into a system. And the act of connecting the chip to the package can take either one of two forms: one is the traditional way that it’s done, which is called wire-bond technology, which means you basically literally connect the chip to the package with a piece of wire. So these wire-bond technologies have been the technologies that have been used, and are currently being used in very high volumes. But as frequencies become faster and faster it’s very difficult to transmit large amounts of information through these individual wires. In that case, people in the semiconductor world have adopted what is called a flip-chip technology, which means the chip is inverted and placed onto a substrate and the interconnector is now all in the substrate, so you’re no longer dealing with thin wires that are connecting the chips, you’re basically inverting the chip and placing it on the substrate and creating a good mechanical, electrical, high-integrity bond between the chip and the substrate. So that is what flip-chip technology is about. Nobody… well I mean a lot of people have talked about using flip-chip technology in photonics, but for a variety of reasons it is currently not broadly employed and POET has spent a lot of time enhancing the capabilities of the technology such that flip-chip technology is an integral part of the interposer. In fact, the interposer works on the premise that chips can be flipped (and) placed onto the interposer and the interposer does what the interposer is supposed to do, which is to provide electrical interconnectivity between all of these chips that are placed on it. So flip-chip of a photonic device is fundamental to the proliferation of our interposer technology into products. So that’s what makes this past announcement last week so foundational to what it is we’re trying to do. In order to make the process of flip-chipping a chip onto an interposer effective there are four critical areas that need to be solved, which is not normally thought about when you’re talking about wire-bond: it has to be mechanically robust, which means it can’t fall or move or jiggle, and we’re talking within micrometers, one-billionth the size of the human hair, that’s the dimensions we’re talking about, so it has to be mechanically robust; it has to be electrically robust, which means the electrical signals that I’m transmitting to this laser cannot have any resistance or high contact associated with it; it has to be optically robust—and this is where a lot of people have tripped up—because photonics devices are finicky and care has to be taken when you’re flip-chipping them such that the integrity of the optical signal coming out of that laser is not effected by the act of flip-chipping, and that took us a while to develop the IP and the capability around how do you flip chip a device and still have it optically be sound?; and the last piece is these devices, once they’re on the interposer, need to be thermally robust. So there are four critical paths that have to be protected and that, again, is what makes this an exciting announcement for us because it has taken us a while to get there but it’s also revolutionary, ground-breaking, and foundational to the interposer technology in terms of solving these issues and being in the position to say we are the first in the industry to do this, the first in the industry to have done this effectively with a DML laser that doesn’t degrade the mechanical, thermal, electrical, optical properties to make it now an effective device that can be used in communications going forward.
(14:19) Suresh: And, finally, these DML lasers that we’re dealing with are probably the smallest photonic devices that have been handled in any form. They are even smaller than VCSELs, physically. So that takes some skill and capability in order to adequately flip-chip and place it on to the interposer. So I think we, as a team, are extremely excited. You know there have been a lot of industry firsts and you hear us talk about many industry firsts on this platform, but this is a big one: to be able to take a really small DML laser die, flip chip it with no detrimental effects, and something that we know is foundational to the proliferation of this platform. So that is important. I just wanted to set the tone for what it is that we’ve talked about.
(15:15) George: Yep!
15:16 Suresh: In a manor where maybe people can understand what we’re talking about and its significance to the company and to our products going forward. So maybe I’ll turn this back to you now George and maybe you could go through some questions that you feel might be front-of-mind to our investors?
(15:35) George: Yeah. And Suresh, the POET message is an ongoing thing that we’ve got here so those kind of descriptions that really go out of their way to explain it make it so much easier. That was great. My first reaction then—even though I have some questions—my first reaction is to put it into some kind of perspective, is what has customer and industry reaction been to this industry first? Give us a sense at home just how big of an industry first this is?
(16:09) Vivek: Yes…
(16:10) Suresh: Vivek why don’t I…
(16:12) Vivek: As Suresh already described why this is meaningful, I’ll give a little bit more colour to it, George. In my experience in about thirty years in this industry, as you go to higher speed, you need to modulate at faster speeds in order to move the signal down the laser line. Here what we’ve done is… in order to do that, in this case in order to increase the speed you use multiple colours of light that you combine together using what we call an optical multiplexer, which is on our interposer. But in order to do that, and these are very narrow beams of light, you need to use lenses and micro-lenses that sit on epoxy and align all these lights. So in order to increase all these speeds you have to do that. What the breakthrough we’ve done here is we have used, this flip-chip uses, traditional semiconductor type of pick-and-place and bonding equipment, which can take the laser, flip-chip it, put it on our optical interposer, very similarly to semiconductor packaging, versus boutique optical packaging, which requires a lot of time, a lot of expensive capital equipment, it’s a slow process so you can’t get cost and scale. Again, we believe that this is a huge breakthrough for our customers, reducing the cost for them in this huge transceiver market for data centers which, as we put out in the press release, is a 2.5 billion market opportunity and still increasing. Then, as we go to 400G, it increases even further… so about 3.5 to 4 billion in a few years. So a huge market opportunity where we can decrease the cost for our customers as well as increase the scale per capital investment of dollars so they can produce more and service the huge data center market that is growing. And 5G markets.
(18:27) George: And Vivek, the customers share that enthusiasm, you know when…
(18:35) Vivek: Absolutely. Yeah, George. We mentioned in previous discussions with you that we have been working with some lead customers in lock-step here. So they have been aware. They saw the potential and now that potential is coming to realization and become a product. And other customers, that were not completely engaged with us, have also seen the benefit of what this optical interposer platform, and developing things like flip-chip DMLs on it, will provide to them. So it’s exciting for our customers too.
(19:11) George: Why is the flip-chip of DMLs important? I know you guys…
(19:18) Vivek: So, as Suresh mentioned, when you go to higher and higher speeds you need to use an external modulator, like a shutter. What we have done is used a DFB laser, which is a high-speed laser at 25G, we have flip-chipped it, which means it’s passively integrated using…meaning you remove that external shutter as a piece part to the whole thing, as well as the cost of putting that piece part in and aligning it, so all that cost is removed, as well as it uses the traditional type of semiconductor packaging versus the boutique and expensive optical packaging techniques.
[20:08, George] And is that the differentiator? Vivek… at the end of the day, that only Poet can do that?
[20:13, VR] So, as far as we know. Again, we don’t know everything that others are doing. But what we have been.. I can just talk from my experience. It is a challenge to do this. But the optical interposer, what it provides is just like a pc board in electronics. Where you put the ICs, resistors, capacitors on it, and you bond it. That’s the type of platform that we have come up with. The waveguides, optical multiplexer, optical demultiplexer, are all laid out on the surface, so these devices go on the surface, which makes it much… ok, relatively speaking, easier to do. Other companies that try to do it… They have to try to dig trenches in silicon and put it deep in, which causes a lot of stress. It reduces the flexibility of.. you know bonding it and aligning it. So yes, I do believe what we have done is very differentiated.
[21:14, George] So Gents, What is the, and Tom I’m not sure if this is your purview here, what’s the market opportunity for this flip chip DML laser?
[21:28 SV] I think I’m going to let Vivek take the full scoping of the market in terms of size but let me just give you a sense of where a DML laser is used. So, a DML laser today, for example, the biggest market opportunity that we are going after is the 100G data center. But beyond the 100G there is the 200G. And its used in 5G networks, either as fronthaul or backhaul. DML lasers are also available in 10G form at higher temperatures, which are again used in 5G networks. And finally, they are also used in what are called long-range communication networks, or LR type networks, which are beyond the data center. They connect data center to data center, as opposed to within a data center. And those are also using 25G DMLs. So, I just want to give you a perspective that, while our initial market entry and the market sizing, and the revenue potentials that we’ve talked about are tied to this 100G application, which is obviously the one we’re going after with our venture partner, SAIC, the applications of both the laser and the technology can quickly proliferate to a much broader set of applications. And, you know, one of the products that we did tape out as part of our multi-product vehicle is starting to look at that. Its going beyond the 100G, and already starting to pre-fetch 200G, LR4, and some other applications where the same laser, the same assembly process are used. So those are… that’s the benefit of the platform. So, I’m going to let it... Let Tom and Vivek work to take you through the sizing of the market in terms of the actual opportunity we’re talking about.
[23:22, TM] I wanted to just jump in, George, and maybe amplify on something we put in the press release. When we talked about the JV and what the JV was about, we were talking about the transceiver market for 100, 200 and 400G optical engines for transceivers. And we put a revenue target on that joint venture of about $250 million in 2025. In the last press release, I said that we believed the opportunity in the same time period for just the 100G segment was about $100 million in the same time period. So, there was some confusion about the difference between 250 [million] and 100 [million]. 100 is only part of the 250. The other two speeds are 200[G] and 400G. Which particularly 400G will start to really climb in the out years. But 100G is the biggest market component today. And I also wanted to make the point that we’re simply talking about transceivers or optical engines being sold specifically by the joint venture. And that’s the number we have put on the revenue expected or the revenue target for the joint venture. That doesn’t cover any other applications that the optical interposer can be utilized for.
[25:07 George] And now my question is irrelevant to any market other than transceivers.
[25;12, TM] That is correct.
[25:19 VR] So just to size the market again… You know maybe there is some repetition, but it’s a market today… Just the 100G space is a 2.5 billion dollar transceiver market opportunity. We play a big role, big portion.. you know, a big, more than 50% of it is in what we call the single mode using the DFP DMLs at 25G. And then in addition 200G, 400. And 400G is rapidly increasing in the next, you know, horizon of 3 to 5 years. So we have a, you know… In 2025 if I was to pick a multi-billion opportunity here which we play a significant role in.
[26:07, George] Ok. That’s saying... That’s really saying something, Vivek. That’s great. We kind of touched on this earlier, so I want to go back to it. I always like outside validation, so customer reaction has been pretty positive, but they’ve known what you’ve been working on. Since the release have you guys heard from outside customers, from the industry in general, or prospective customers about this achievement and their reaction.
[26:34 VR] So we have shared with customers. We are setting up meetings now to discuss further with them. But clearly there has been a positive feedback from the customer base, knowing that we have achieved something that will provide a much better cost structure to them. In a market which is growing but the prices are coming down, so they are struggling to maintain margins. So, yes, to answer your question George, yes. So, there is significant customer interest. We’ll have to go into details, discussing with them... you know and providing samples and getting and winning their business.
[27:20, George] What kind of cost savings are you delivering?
[VR] So as we said, just for 100G CWDM, that’s the lowest speed where the prices are the most suppressed, okay so let me start with that. We can deliver 25% to 40% cost savings to our customers where they can either reduce their end pricing and still make money or increase their margins… so both situations. Over time the prices are reducing rapidly in the 100G. But then if you move to 200G CWDM-4, you know the prices are slightly better. Our cost structure, what we provide as a cost is not too much higher. So again, there is a big opportunity there. And there are other emerging formats as Suresh mentioned in 5G networks. In front haul, mid-haul, where there, even instead of 4, CWDM-4, you have CDWM-6. So, you have 6 of these lasers in different colors side-by-side. The benefit of what we are providing really escalates at those points, you know. So, when you have multiple lasers being flip-chip attached on the interposer, we provide a much higher value.
[28:38 George] We may have lost Suresh’s voice there, but we got most of that… sorry, Vivek’s voice. Gents the Chinese have a saying, and maybe I’ll bring this to you Suresh, the Chinese have a saying “Be careful what you wish for” Do you have the resources if you really achieve everything you want to achieve here. Do you have the resources to handle the new customers, the onslaught of what potentially could come of this over time?
[29:10, SV] Yeah, I believe what we’re doing is… Look I think in order to succeed we don’t want to overreach and overstretch. So, we’ve been very articulate about our roadmap. And we’re sticking to that. We do think that forms a good underpinning to our business plan. I think Vivek has presented this at the AGM in August and we’re executing to that plan. So, the plan allows us to systematically walk before you can run. I think we… well as Vivek said, there’s the technology and the interposer has always had promise with customers. They look at what we’re doing and they say “these guys are working on a different, unique, differentiated, easy-to-understand approach to how photonic devices are assembled and packaged.” But the burden of proof has always been on us to demonstrate across a couple of different applications. And we are slowly getting through these major system-level burdens of proof. For example, “can you do a DML?” Well now we say we can. “Could you do a CW laser with good coupling efficiency?” I think a week ago we presented that we could do that. So, we’re slowly getting past these critical burden of proof requirements we had on us. Not at the component level… we had done that a year ago. But now at an assembled, system level. You know there are these hurdles you’ve got to get through. And so these things that we are releasing now are significant advances to our capability because critical burdens of proof at a system level are now getting solved. And that allows us to get one product out and then we would proliferate beyond that.
[31:10 SV] That’s where also the JV comes in. And we will probably be sharing more information about that to our shareholders. We started to get the management team of the JV put in place, hiring people into the JV. And you know the initial capital expenses that are needed to establish the lab and the fab are starting to be put in place. So its not just Poet. We’re getting to have the benefit of this JV come in and support, if you will, the building, qualification and the ramping and scaling of these solutions that Poet has been developing over this period of time. So I think we are being deliberate. We don’t have infinite resources. We cannot pick up every shiny rock along the way. But we have a road map and I think its critical that we focus on executing to that road map which we believe will result in success for us, the company, the shareholders, as well as potential investors.
[32:15, VR] [garble] We have built, brought in, key, experienced people to the team. To provide expertise at, of all different types, levels that is needed to do. And we are also partnering rather than doing and making everything ourselves. We are partnering and strategically partnering.
[37:37 George] Suresh, I like by the way that you said that. And its funny. I use that example of the shiny rock. You can’t pick up every shiny rock, otherwise you’ll drop them all at the end of the day. And I like that you have the sense to say that we’re not going to try and grab everything. That’s good to actually hear. Some shareholders might think, “You’re not going after every opportunity?” And you can’t. Success actually comes from leaving some opportunities on the table and focusing on the right ones. I’d be remiss if I didn’t ask guys… Devil’s advocate. What’s your Achilles’ heel? Not that that means its fatal. But every company has its Achilles’ heel, something that keeps it up at night. I don’t care if its Coca-Cola, Google, Agoracom or Poet technologies. There’s always some sort of Achilles’ Heel out there. What would it be for you guys?
[33:30, VR] So the way I see today what if you can refer to it as the Achilles’ Heel is that there is a certain process of winning business. We do need to sample customers, and customers do need to design it in. And these are not commodities, right. Although there are certain standards, it still takes time and collaboration with customers to design in our products. It has to go through a qualification cycle. There are certain qualification times which are industry standard you have to meet. You know you can’t say.. there’s no shortcut to it. And it’s important for that. So, I would say at least from my point of view, we have overcome most of the, if not all, and we’ll keep announcing and sharing with you and investors as we make these achievements in our platform. But you know ultimately, you know this time we cannot shrink that time from the qualification and getting revenue and business. So, I think that is really what… the situation today.
[34:49, George] Which probably explains why you guys are a pretty calm group at the end of the day. You realize that there’s a process here, and that you’ve just got to go through, so you’re not… You know a lot of small caps sometimes will try to jump up and down and talk about all the shiny rocks in the catch but I like that approach.
[35:02] I think Vivek hit the nail on the head back in the AGM in August. We’re getting to a point where…Look I think having done the research, moved to development, and then go to products, and then you know the number of gotchas go away and schedules become a little more deterministic. And I think that’s the phase we’re on. We, you know… there is a certain time period unfortunately to roll these things out. And if there was an Achilles’ heel for Poet, I mean we are fabless today, and so we are dependent on a lot of partners working with us on lock step, and some of those schedules and timelines are, you know, we wish they could have been faster, but they follow a certain lead time that is consistent with what the industry uses and we have to live with. So, I think… but the fact of the matter is that the number of, I guess, issues that we have to solve, or, you know, I used to call it, that’s whack-a-mole. You know early in research there’s so many issues that there are these moles that keep coming up and you’ve got to keep whacking them right? And in software they call it the bug rate. The rate at which you solve bugs. And you know that piece is now, to be, you know, it’s mostly done, and we’re down in this deterministic development phase which is invigorating for us. I mean we’re seeing obviously, you know, the light if you will, at the end of the tunnel coming out the other side of our interposers and usable light can be used for products.
[36:50, George] Its Dec 21st as we’re recording this and talking about this. So the year’s almost done. You’ve achieved a new 52-week high today. So congratulations there. What’s next for Poet? Is the year pretty much done? Is there more that can happen for the year. More importantly what’s next for the company on its roadmap.
[37:17, SV] Oh the year’s not done, oh, my God, no. We’ve got work to do. Look, for us it’s a real focus on customer acquisition and product release. You know that’s what I think our focus is going to be for the next six months. I think we’re… you know, we’ve got a package of information now with a lot of these burdens-of-proofs now behind us, that Vivek and team are taking to the customers, you know, both the ones that we’ve engaged with previously as well as new ones. So, our focus over the next six months is unchanged from what it was back in June. I think the, you know…we were deliberate about our roadmap. We were deliberate about our communications. We felt that it was important to communicate with transparency. But also the timing of the communications needs to be consistent with our degree of comfort. And you know we want to be credible with the information that we put out there. And so I think that as you see us now come out with a cadence of releases, it is… there’s an underpinning of real data that goes behind it and that we can stand on. And we feel comfortable about that. So we’re going to just keep going down that path. Nothing changes for us as a company just because we had a good day in the market. You know we expect that as we execute there are going to be many of these days. And I think… I hope this becomes a norm rather than an exception.