Here is a repost of ITTR’s beginning with my ending, all in one post.

Rogue

Suresh Venkatesan: Thanks for the introduction, Alex, and thanks for opening up your platform for us to make this presentation. Good morning, good afternoon, good evening—depending where in the world you are—to those watching. Again, thanks for your time.

Over the next thirty minutes I’d like to do a good overview of POET—POET Technologies—what we’re about, why we’re doing what we’re doing, where are we in our commercialization process, and why we believe the company is poised for growth and is a good investment play for the folks interested. 

The presentation will cover four specific areas. I will spend a bit of time providing some background on photonics in general and talking about POET Technologies at a high level… how we’re sort of servicing specific aspects of this photonics market. I will spend a bit of time on: strategy, markets, and products. Then, our own execution of that strategy, which is operations, growth, and revenue plans. And then, finally, the investment highlights and why we believe POET is a great opportunity. 

Photonics is an enabling technology. Photonics is the science, if you will, or the technology behind generating and harnessing light. Light is, of course, the fastest moving electromagnetic wave in the world. Photonics typically uses words like: lasers, optics, fiber-optics, and they’re used in a variety of different fields. Most photonic applications require a seamless integration of electronics and a variety of photonics components to create ultimately what would be the optical device, whether that device is a transceiver or a sensor, ultimately there is a requirement to integrate various types of components. We classify them in broad categories: like photonics, that involves photons, so those are lasers, detectors, and so on; electronics, of course, to control the photonics’ components; and then we’ve got optics, which are like mirrors, lenses, etc. So most photonics applications typically include, or provide, a marriage across these three component sets. 

Photonics, in general, has found its application in the semiconductor industry across three largely distinct verticals. And I would say it is application specific. The first major application of this technology was for communication. This started at AT&T and Bell Labs, and then over the 2000s, communications has been the big linchpin, if you will, for the applications of photonics. Here, we’re talking data and telecommunications, data centers, 5G, as well as the edge computing networks that are currently being deployed with the 5G infrastructure for machine to machine communications, IoT, etc. 

The other big class of photonics applications has been sensing, and that has been around for a while as well. Of course there are new technologies that are being evolved in the sensing space: medical, autonomous vehicles… But prior to that it’s what we use in our cellphones today: time-of-flight based sensing devices. So that’s been another major application of photonics. There are industrial applications, that I’m not talking about, like laser welding and so on. But, broadly, in terms of the really high volume segments, these are the ones that we talk about. 

Then, there’s this burgeoning field of computing, which is using photonics to create compute-engines for artificial intelligence. And there are a number of companies starting to put together architectures, as well as solutions, toward neuromorphic computing.

So those three, from our perspective—from POET’s perspective—provide the biggest opportunity for our technology. So those are, broadly, communications, computing, and sensing. And we think that this global market is about half-a-trillion dollars, 500 billion dollars in 2019, growing at about a 7% rate, to an extremely large market segment. And we believe that these are application segments that POET, with its somewhat versatile as well as ubiquitous packaging technology approach for photonics, can play a part in.

Now, what is the problem statement? Why are we talking about our technology? Why are we talking about the optical interposer? The issue with photonics has been that conventional approaches to assembling photonics devices, while at the end of the day providing good performance, are expensive in both capital as well as labour. So the really high volume applications of photonics, they struggle in regards to margins because you’re effectively stuck with a capital structure, and a labour structure, that does not scale at volume the way that semiconductor technology typically does. These components and subassemblies are assembled one at a time. They are not assembled in volume, and each assembly requires optimizations of the signals, that we call active alignment, in the field. And there are many, many subassemblies involved. As a consequence, there is no economy of scale: there is a one-to-one relationship between input and output. The analogy I used once was, you can either make an ice cube with a single ice-cube mold, in which case you’re making ice one at a time… if you’ve got thousands to do, you have to do it a thousand times… versus creating a tray, and you fill it with water and you get 100 ice cubes at a time. And that is the difference between what POET does with wafer scale, where we do many, many at a time, versus what is conventionally done which is singular assembly. 

The problem with singular assembly is it uses multiple subassemblies, and each of those subassemblies uses multiple components. So you basically have a factor of ten increase in the number of things that you have to handle to create an assembly. With POET that is dramatically reduced. We reduce capital costs by a factor of ten. We reduce the bill of materials by a factor of ten. And we transitioned this entire model from an extremely labour intensive model to something that is wafer scale, so it’s a wafer scale test platform, a wafer scale assembly platform, and basically a wafer scale integration platform that we call the optical interposer. We believe this is an unprecedented advancement in terms of scale as well as cost disruption as it relates to photonics. 

So how does POET solve this problem? Like we talked about, the solution is moving to wafer scale. It is full integration of multiple active components with passive alignment at wafer scale and we use conventional semiconductor assembly technologies to do that. And per my previous analogy of the ice cubes in a tray, we have an extremely non-linear relationship… you know, you fill the tray once with water, you freeze it, and you get 100 ice cubes. We do the same thing with our interposer: we essentially assemble our optical engines at a wafer scale integrated form factor and that dramatically simplifies packaging, lowers bills of material, uses highly automated wafer scale manufacturing techniques. And perhaps what is also missed is we have the densest, smallest form factor, which means…conventionally if you could fit a single engine in the space available to you, you now can fit up to four. It provides you some alternative architectures that you previously couldn’t dream about doing. So it does open up the space of applications because of that kind of density and form factor that we offer. Of course we do that while maintaining very good electrical as well as optical performance. 

We’ve been able to deploy this kind of approach, and we’ve created the world’s smallest and lowest cost 100G optical engine, scalable to 200G… it’s the same engine that can expand that range. And it includes all our active and passive photonics devices. We’ll talk a little about what that product looks like in a bit, but this is the way POET is solving this particular problem in the industry which is, hey, we all want photonics, but we want it really cheap, and we want it in really high volumes, and today’s technology really just can’t do that for you. What POET does is provide a very, very versatile platform that enables for that problem to be effectively solved.

POET’s been at this for about four years. We started actively working on the optical interposer in about 2017. It’s now 2021 and we’re getting to the point of rolling out products. It’s been a multi-year development project, with multiple interrelated projects mingled in. I think the true scale of what it is that we’re trying to solve… we’re trying to do something for the first time that nobody’s done before… you’ve got to start with architecture and end with a product. And you have to do everything in the middle. We’ve been focused on working out the kinks on the interposer, all of the passive device designs, you have to create new active designs that are flip-chip and compatible with our interposer, and then solve all the assembly and back-end issues that exist, and then ultimately do the product design. Mid last year we announced we were pretty much done with the technology development phase, and ensured that they were validated, and we moved toward the optical engine product design, which is solving the last of the issues that’s associated with the product. We believe this is epic in scale and time, especially with the amount of capital we’ve deployed within the company. We’ve also achieved proof of concept of these optical engines when we did these our pre-alpha samples toward the end of last year into the first quarter of this year. We’ve taken all the learning and we’re in the process of putting together products that can be sampled to customers in a more-or-less fully functional form that has been pre-validated. 

So why go through all this? What are we going to get at the end of the day? The benefits of using POET’s optical interposer add up to a truly disruptive entry, especially in large scale, high volume, photonics markets. We believe we can lower the end module cost from anywhere to 20-40%, and in some cases it is higher than that. It really depends on the applications that you are going after and the volumes that that particular application asks for. Our CAPEX investment is easily 10X than what would otherwise be able to be achieved. Because we do a chip scale packaging, and we have extremely good thermal characteristics associated with the interposer, we believe we can effect a lower power, at the end of the day, at the module level, and most of that comes with overall better RF and overall better thermal as it’s associated with the interposer. Wafer scale assembly and test, you know we have 100… you know I think our current architecture has about 350 optical engines on a single 8” wafer. So you’re making 350 ice cubes at a time as opposed to one… a very simple analogy of how big an effect this can be. It’s a planar architecture, so it provides a lot of flexibility. We can address applications from single channel, to four channel, to eight channel, to greater number of channels, far more easily than any other technology out there. In fact, the greater the number of channels, the greater the advantage that POET has. That is really important as we are migrating from 400(G) to 800(G) and there’s a need, when you go to co-packaged optics, to have eight, sixteen, thirty-two channels, and we believe that we provide that flexibility advantage.

We also have—unlike silicon photonics which is more traditional in its monolithically integrated aspects—we have the flexibility to address other applications in PON (passive optical networks) and so on and so forth. It’s not our immediate focus, but we do have that flexibility because of the architecture that we have in the technology. Of course it’s a platform… we’ve discussed this before… it’s like Lego blocks, you put it together in various forms which provides the versatility to address many, many applications that we believe we can go after with this technology.

[slide 10] If you were going to look at the optical interposer in terms of POET and how it compares to traditional micro-optics assembly, these are things like what InnoLight, or what Finisar, II-VI, has been doing for a while. You could also compare it to a fully integrated, or monolithic integrated, kind of silicon photonics solution (Intel and Luxtera indicated in this category on the slide), and you can see across a variety of different categories we believe are important to assembly and scale of optical engines, POET definitely shows an advantage. Our hybrid integration platform is what it is: it’s hybrid. It’s not intended to be monolithic. But in my experience of over 20+ years in the semiconductor industry, if you can come up with a good hybrid integration platform, that allows for the integration of hybrid components, those technologies tend to win out in the end. And we’ve seen that in the electronics world with multi-chip modules that include GaAs and SOI technology and baseline CMOS that effectively all co-integrated with a low cost, high performance, high throughput, high scale, packaging platform, and that’s what POET provides to the world of photonics.

So a little about our strategy, markets, and products. Our business strategy, number one, is to support Super Photonics as an independent company to drive growth in optical transceivers, and deliver maximum cash flow to our partners. Number two, we’ll continue to engage with industry leaders and incumbents to design, develop, and sell devices based on the optical interposer. So those two come hand-in-hand. POET goes to market, brings customers in, Super Photonics, as our joint venture, basically is an order fulfillment… an organization that fulfills the orders that we achieve. (3) We want to exploit the localization imperative in China. We believe that there is a large, and burgeoning, and growing market, not just in data communications, but in other fields of optics and photonics, within China. And we believe the localization imperative provides POET, Super Photonics, and SAIC our partner, with a very unique opportunity to seek both an organic and inorganic growth. (4) We want to form additional partnerships in other target sectors to establish fabrication and sales operations globally. (5) We want to pursue complementary strategic alliance or acquisition opportunities for inorganic growth. And then finally (6) we want to explore technology licensing opportunities for growth in non-target sectors. We believe, again, the platform is very versatile. We’re not going to be able to address every opportunity out there within POET, but if it is not strategic to POET, but we think there is value to be had, we would be considering augmenting our business model with licensing opportunities down stream. So that’s kind of in a nutshell our six-point strategy from a business perspective as we look forward to putting products out, how do we intend to grow and monetize this company. 

[18:50] Our target market is initially in optical transceivers… You know even as 400G has been critically important there’s been some delays in its deployment largely due to technical and cost reasons. The 100 and 200G segments continue to be large and attractive markets. We’ve not seen any major disruption in the cost structure in those applications. And we believe we come in with a fairly step function disruption in the cost structure for these applications in 100 and 200. And so, our focus and priority and strategy continues to be, provide over this next year solutions for 100,200, and 400G. And then we use that kind of as the launching pad for going into other applications. And these markets themselves are fairly large in and of their own right. Over 6 billion dollars of potential SAM for Poet.

[19:55]  So our commercialization roadmap, our current focus for 2021/2022 is 100/200/400G data communications. So that is our focus to seed the market, to deploy our first product solutions. We’ve been, like I said, been in the development phase for a long time, whether it be technology elements or platform elements. Now we believe we’re going to be able to put products out, fully tested and validated in customer’s hands shortly. We’re in a competitive position in terms of scale and cost in the data communications market. And its also an established market with an extremely strong growth and very high volume. So our first and foremost, our focus is, get our products designed in for 100/200/400.

[20:40] The near term opportunities thereafter, are expansion. So, once you have that beachhead, you expand. So that then is second-generation 400/800, co-packaged optics, 5G. We would have some proprietary high channel count solutions based on our integrated platform, as well as deploy some high density optical interconnect solutions for remote light sources. So that is kind of… expand. And then finally, we would address other verticals. AI computing. We’ve talked about that. Autonomous vehicles that require sensing, as well as health. And we believe that Poet, because of its ubiquitous nature and versatility has some very, very interesting possibilities and applications in the burgeoning health arena, where point-of-use type devices are critically important.

[21:44] A bit about our development cycle. We usually talk about things in the pre-alpha/alpha, and so on. I suggest a bit of explanation, as we move towards product development versus technology development. Pre-alpha prototypes, these are validation. These are primarily tested at Poet, data shared with customers, and it validates the feasibility of the design and the platform. We’ve completed our pre-alpha for our 100G early this year in Q1, as well as towards the end of last year. Alpha is when the prototypes are tested at Poet and then handed over to customers for them to beat on and shake up. These would meet most key customers specs. Our internal qualification would begin with these alpha prototypes, and I say “qualification”, in this case, it’s not component qualification, it’s the [entire] engine qualification. Then we move to beta. At the time we get beta, it meets all the specifications. So, at alpha its possible that certain things are slightly off spec. And those would create some actions, that would create modifications to the design. It gets to beta. At the time we are at beta, basically it meets all customer specifications, and it is suitable for the customer to qualify. Finally, as the customer is qualifying, they would be a slow ramp of production. So, these beta prototypes usually go into production and start shipping in low volumes, and we also propagate the solution across a broader customer base. So, we are currently in this phase where we’re in the process of creating alpha prototypes. We have previously announced a couple of months of delay because of the shortage that people have talked about in the chip world, that has caused delays in the delivery of wafers. I am happy to say that those have now been delivered and we’re now actively in the process of starting to build some of these prototypes that get in the hands of our customers. We don’t believe that there is any material change to our revenue plans, but I think its important to acknowledge that we’ve had some delays as a consequence of the global shortage of semiconductors we’re seeing over the past year.

[24:15] So what does that mean to our roadmap. So again, our roadmap is largely unchanged, but it does reflect a delay as a consequence of, not our execution, its just a consequence of the way the industry has been able to respond to some of the issues that the pandemic has caused. We expect, like I said, to have alpha samples of our marquis products in the 100/200G regime over the course of the next few months. I think it’s imminently in the process now of being assembled, and then we will go through testing, and then engagement with customers. It doesn’t fundamentally change our production timelines which would be towards the end of this year into early 2022. And similarly, we’ve had some delays across the board, but I think if you compare this to our previous roadmap there’s not a material delay. We are still on track. And we believe that across 100, 200 and 400 we will be in a position to get alpha or beta depending on the type of product by the end of this year, on our initial focus of the products.

[25:36] The 100/200G datacom solutions, basically there are 3 products times 2 form factors. So, we either have fully integrated transmit/receive, or transmit only or receive only. These are products that use DML laser solutions that were I think were the first in the world, and I think the only person in the world to flip-chip DMLs at wafer scale. I’ve not seen anybody else do that. We have made this platform now compatible with CWDM, which is coarse wavelength division multiplexing as well as LAN-WDM which is denser, to provide LR4 solutions. So, we expect this set of products to be CWDM LR4 data centers. We also have some custom solutions like I talked about before. Because of our form factor there are some potentials for custom solutions where 2 or 4 of these can be integrated into a single module. As well as solutions for 5G interconnectivity. As you can see on the right, some proof points. As we’ve gone through the pre-alpha phase, where we’ve got excellent eyes, which is a prerequisite for good performance. In the networking or transmission world it’s called an eye diagram. So if the eye is wide open that means you can pass signals through without any distortion. And we’ve been able to demonstrate that with 100 as well as 200G with the lasers that we’re currently utilizing from SAIC.

[27:18] For 400G datacom solutions, our solution provides the interposer, the lasers, the mux, the demux, especially for FR4. And we would partner to get a 400G modulator. And these products, again, use wafer scale hybrid integrated photonics, low-loss transmission and coupling. In this case, we would be using continuous wave lasers up to 55mW, that we have already designed and are going through the beta phase of these 55mW lasers. And they would be coupled with a silicon photonics modulator. The immediate applications for this are in the DR1/DR4 ,FR4 and also DR8 which is an 8 channel application for data centers, and then of course subsequently, co-packaged optics.

[28:10] Our strategy in manufacturing is asset light and capex light. So, we use a high volume facility, Silterra, to manufacture our optical interposer wafers. And they’ve got plenty of capacity, 30,000 wafers a month of capacity. So we don’t believe we are going to be capacity constrained at that foundry. We of course work with SAIC which is our high-volume III-V semiconductor foundry. So they provide our lasers, and PDs and the monitor PDs, and the active devices that are hybrid integrated onto our platform. They are the world’s largest III-V compound semiconductor manufacturer today. They provide the kind of scale that we require. And of course, in terms of go to market, order fulfillment, we have created a joint venture between Poet and SAIC and this joint venture creates the optical engines that ultimately are sold to our customer base. In this joint venture, SAIC will invest the capex in order to be able to scale manufacturing. And we have the ability to leverage this joint venture to address the large localization initiative as well as the local markets in China. So, we believe that this strategy that we have deployed for manufacturing between Poet, SAIC and Super Photonics, effectively constitutes a pseudo-vertically integrated model. And that provides unparalleled cost efficiency, as opposed to having margin stacking when we’re dealing with multiple potential vendors.

[29:50] In terms of execution, we’ve augmented our management team. The executive team is myself, Vivek and Tom Mika that all of you have met before. On the engineering an operations team we’ve got Ed Cornejo, VP of product marketing, Dan Meerovich, VP of product engineering, we recently announced the addition of Dr. Mo, in Asia, And then of course we have James Lee and Kevin Barnes. So these are our engineering and operations teams. We’ve got 35 employees, 30+ engineers, and we’re growing. There’s a growth plan for this year to meet the demand. Not only inside of Poet, but also inside our joint venture of Super Photonics.

[30:35] Our development and manufacturing is global. Toronto, Canada is our corporate headquarters. Allentown, USA is optical engine product development. In Singapore, we do all our interposer development and testing. And Shenzhen, China is our applications engineering team, as well as some evaluation for customer engagement activities. Xiamen, China is our joint venture for assembly, test, and go to market sales. And Kulim, Malaysia is Silterra for our interposer manufacturing. So we do have a fairly global development and manufacturing footprint that allows us to leverage the unique skills that each region has to offer, while maintaining a footprint in the two largest markets, North America as well as China.

[31:24] The supply chain is licensed to produce optical engines for the transceiver. So, of course, Silterra will be producing our optical interposer based on Poet’s device and process IP. They are of course licensed to manufacture these interposers for only Poet, including using our consigned equipment there. Our Poet device and design IP goes to the III-V manufacturers. Again, they are only licensed to manufacture for Poet and the optical interposer. And these are the active devices for the optical engines. Currently we are working with Almae in France, Denselight in Singapore, and we’re really bringing on SAIC as the big player here to provide most of the active devices that we would require, at least in the DML 100/200G regime. And then for the electronics devices, that go with the optical engines, these are merchant market vendors. So these are vendor owned IP that we would buy and include as part of the optical engine. Or at least include as part of our evaluation board and validation strategy. This goes to super photonics that does the marketing and sales, assembly and test, and effectively order fulfilment, of the orders that Poet enables as part of our deployment strategy.

[32:52] [SLIDE shows contracts:2, Finalizing Product Plan: 3, Requested Samples: 6, Active discussion: 14, Targeted: 28, Target Company Names: Google, Arista, ADVA, Cloudlight, Delta, Eoptolink, Facebook, Source Photonics, Hengtong Group, Amphenol, Nokia, Hisense, Jabil, Molex, Xgiga, Alibaba, Huawei, Cisco, Hisilicon, Accelink, AWS, Juniper] We’re happy to have a lot of customer engagements across multiple verticals. Really we’ve got a couple of contracts, we’re finalizing plans with 3 others. And we’ve got basically a pipeline of customers that are waiting to evaluate our alpha samples. At the end of the day, as Vivek always tells me, there’s just been so much new stuff in optics that just hasn’t kind of worked out. That there are a lot of people that are gun shy, and its up to us to deliver proof of performance. We’re actively working towards doing that. So, we are happy to have this customer engagement. Of course, you know that means we have to deliver, which we’re putting forth execution plans to do. But the engagement has been good. The interest has been great. And we believe that once we have samples in the hands of a few customers that are fully validated, that we believe that we can really grow this capability, and grow our business along with it.

[33:58] [SLIDE: Super Photonics revenue estimates, Year1 1..2M, Y2 15..25M, Y3 90..150M, Y4 140..240M, 180..300M] The range of forecasted revenue for super photonics.. Of course it would initially start with 100, 200 and then migrate onto 400G engines. We have forecasted this to be in the range from 180 to 300 million dollars, depending on the penetration rates that we would achieve. And this is specifically again for the immediate focus of products.  We would expect that as we get into next year we have additional market verticals and additional applications that would augment this revenue. But we believe that this revenue is reasonable for us to be able to achieve out of Super Photonics which provides the underlying value proposition to Poet as well with its technology.

[34:50] The overall scale of our business, if you really include all of the different verticals that we have the opportunity to penetrate. I mean we’re talking initially about the first vertical, transceivers for datacom, and specifically 100/200/400. But then you’ve got other verticals in 5G, co-packaged optics, as well as optical computing, sensing and some of the other applications. And we believe that this opportunity for Poet can be as large as 1 billion dollars, provided we penetrate all of these verticals the way that we’re thinking we can. And we believe the platform can enable us to penetrate it. But, of course, everything starts with a seed. And our seed is the initial focus in business and transceivers for datacom.

[35:42] So I think to conclude, Poet represents a multi-billion high-growth market opportunity, first and foremost. We’ve got a robust product development pipeline with a very strong IP position. We work with a very high margin business. Of course, the cost structure associated with the optical interposer itself provides a disruption. And that enables us to keep sustainable cost advantages. We get a disruption and then the standard semiconductor technology processes provide these cost savings as volumes increase. We’ve got unique technology, for mass market commercialization as well as something that is versatile, that can approach solutions that other more integrated technologies cannot approach. It’s a paradigm shift in the way you think about how photonics devices ought to be assembled. And we believe that we can effectively be a de facto standard for how people ought to think about architecting photonics engines and then the future. And of course we have the ability to leverage a lot of scale with our existing joint venture. So, I think with that kind of a backdrop to potential for the business, imminently today we can declare that our capabilities are proven in terms of the interposer itself. We are in product development with several customers. Alpha prototypes are weeks away. I think we wanted to have these end of March, and we’ve declared that we’ve had some delays. But we’re now weeks away. Beta prototypes a few months from then. The design funnel is filling. We continue to have good engagement and we continue to have customers pulling us in directions that we believe we can support them. And we are gearing up for manufacturing and volume. I think we’re starting to put the equipment into Super Photonics now. First, the first of a kind. Then as demand fills, the second of a kind. The team is being built to deliver. So we believe that we’re kind of poised at a cusp in terms of the realization of commercialization from a company that over the past several years has been systematically building the foundation to get us to this point in time. So that’s what I wanted to discuss today with you guys. Thanks for your time.