Poet Technologies Inc (POET)

[prototype_101]

Here's the bulk of Tom Mika's recent presentation (no Q&A or Mika's remarks on our new hires). All mistakes are mine. Told myself I no longer had time for this, but the recent share price drop made me sad enough to do it anyways. Hope it proves useful to some of us longs.

LD Micro Virtual Conference

March 4th, 2020



Thomas Mika (01:18): So, what’s happening at POET Technologies? In short: a lot. But, first, please keep in mind we’re a development-stage company in a competitive sector, but one that rewards innovation and disruption handsomely and in which extreme rapid growth is possible. Our main news is that we’ve validated the performance of the POET Optical Interposer with a highly regarded customer which has recognized the potentially disruptive solution that it represents. Now that proof of concept is behind us we are in discussions with them and others to design the optical interposer into their products. What has given us new life is the sale of the DenseLight fab in Singapore. That has put both cash on our balance sheet and reduced our cash burn. We have opened—or are in the process of opening—operations in Singapore and in Allentown, Pennsylvania. And, finally, we are enjoying an uptick in our share price, which we believe puts us in the territory where we might see some conversion of outstanding warrants. And I truly hope my presentation today doesn’t take that stock price down (insert sad trombone noise).



(One note from slide three: “Added 9 senior execs and senior engineers to our post-divestiture core team of 16 in the past 3 months bringing our headcount to 25”).



(2:41): For those of you who are new to POET, I just wanted to quickly review with you the importance of the technology that I am presenting today. Virtually all long-distance communications are carried on global fibre optics networks deep on the ocean’s floor or underground. In addition, with few exceptions, all data transmissions both within the cloud—that is within the data centres here on Earth—and between the data centres to whatever connection point, for the last mile or five miles, are carried on fibre optics networks. One essential element that allows this to happen is the optical transceiver. Today, these devices plug into the ports of network switches and storage units, carrying on-demand video to your home and data to your office. Optical transceivers sit at both ends of an optical fibre. There are thousands of them in a typical data centre and even more in a mega data centre, which companies like Facebook, Microsoft, Amazon, Google and others are building at an astounding rate. Inside the optical transceiver is something called an optical engine and that’s what POET is producing. Optical engines include the devices that convert electronic signals into light signals and back again. Optical engines include such elements as lasers, detectors, waveguides, modulators, multiplexers and de-multiplexers, which are all so-called either active or passive devices. I won’t go into a description of all of these and how they operate because it is both outside the area of my expertise and beyond the short period of time that we have, but I can say that this is the absolutely essential device that allows this communication to happen. It is the only device that converts electronic digital signals into light waves.



(4:47) So here is what POET has invented. We call it the Optical Interposer, and it is a platform technology, that allows all the devices in an optical engine to be integrated, which means they all work together seamlessly and they’re manufactured together at wafer-level. Wafer-level processing is what drives down cost. I believe we are the first to demonstrate the assembly of all the components of certain types of optical engines fully at wafer-level and at the same time remain compatible with standard semiconductor processing technologies, usually referred to as CMOS. You’ll notice that on this platform the optical element and certain other electronic devices can be placed on the same layer (slide 6) and they’re using those lines that are embedded in the interposer for high speed communication between the devices and among the devices. On the right of this slide is a schematic of the optical interposer itself, showing some of its features. It is a device that really exists and we’ve spent a fair amount of time developing.



(6:05) As our initial application we selected the market of optical transceivers in segments where our technology makes a material difference. Ethernet on optical fibres is one of the fastest growth segments in data communications. It is a large market, about four billion, and the next generation is still being designed. We’re talking, of course, about 400G transceivers. Now there are many technical challenges for any developers of optical engines for 400G transceivers, but this segment is one that we believe we can penetrate, especially with our embedded mux/de-mux, self-alignment, and the ability to assemble, test, burn-in, and package all at wafer-level. Our second opportunity is 100G. Now this one is more straightforward in a technical sense, but everyone is established already in that market so it will take both the benefit of the interposer, which should bring higher margins to module producers, and the ability to scale rapidly. And I can tell you we are working on both aspects of these challenges.



(7:27) I wanted to go back to the prior slide for just a moment (slide 6). So I wanted to point out what our unique advantages are in this platform. First are waveguides, which include multiplexers and de-multiplexers that are not separate from the platform, they are actually built in to the platform. And they perform in a way that facilitates the placement and the alignment of the active devices, such as the lasers and photo-detectors, automatically without having to test each placement. Second, every step in the process of constructing an optical engine is completely compatible with CMOS processes. Today no other company that we know of has an embedded de-multiplexing device in their optical engine: they’re all built as separate devices, are not CMOS compatible, and must be placed, aligned, and tested one at a time. Third, the elegant design, assembly, scheme, and features of the optical engine give it the ability to be adapted quickly to various devices. We can use the same platform, for example, substituting four 10G DML lasers to produce a 40G engine with four 25G DML lasers for a 100G engine. This flexibility allows the platform to address a wide variety of applications and to give manufacturers of products the ability to span multiple product generations with minimal additional design and development costs.



(9:30) I’m going to skip this slide because we’ve already covered it (Slide 7). These are large markets that we are attacking with good opportunities. I did want to spend some time on how a POET optical engine differs from other optical engines. In the CWDM DR4 space, which is where we’re producing devices, which we believe will become a de facto standard at least in China, about 60% of the current market for these devices uses conventional methods of assembly and test—that’s in respect to the 100G engine. In conventional assembly devices are placed and must be optically aligned one at a time. This is done with massive amounts of machinery and labour. At every placement the possibility of device failure exists, which means a lot of partially or fully completed devices are scrapped. Contrast that with the POET optical engine, which is built 500 at a time on a single 8” wafer. We substitute all those potential points of failure with a single device that is tested and known to be good. Even the fibre alignment needs no lenses, which are common in both conventional assemblies and the silicon photonics-based modules from Intel and Cisco.



(11:05) Although reversed in orientation (Slide 10), I wanted you to see how big a difference this really makes. The control electronics for the lasers, detectors, modulators are on the printed circuit board on the right—in the green area they are the black devices. At present we do not include these in an optical engine, although they could be and they’ll have to be when we reach speeds of more than 400G per second. At that point those devices may need to utilize the high speed metal traces that are built in to the POET Optical Interposer platform. On the left are the gold barrels that contain the components of the optical engine. In this case they’ve been divided between the transmit portion, or the TOSA, and the receive portion, or the ROSA. In the bottom half of the slide the POET Optical Engine replaces both with a single device, smaller in size, and lower in cost. This demonstrates why module manufactures are interested in POET’s approach.



(12:05) I’m not going to go through the details on this slide (Slide 11). I’ve included it to illustrate the development of the optical interposer involved many separate but interrelated developments. We are now moving rapidly from proving that the concept works to the design-in and qualification phases with customers. By the way, the set expectation is design phases usually lasts six to nine months or longer after which there is, depending on the location of the customer, a period of qualification of the new designs, usually lasting another six to nine months. That’s why we said previously that we wont expect to see product revenue until next year, that is 2021. During the year we expect what is called non-recurring engineering revenues, however our operating plan takes us well into 2021 with no revenues actually planned—that’s not to say we won’t have them, it’s just to say our operating plan is not showing any in our plan so that we can extend our cash.



(13:15) As a platform technology the optical interposer can be effectively applied to photonics integration at several levels and multiple applications, including those included here (Data Center Market, Telecommunications including 5G, IOT, Industrial Sensing, LIDAR, and Co-packaged Optics—Slide 12). We believe we are in a very good position to penetrate both the data center market and the market for co-packaged optics (ASICs: switches, graphics generators, microprocessors) as long as we have the right partners and we have so far.

https://agoracom.com/ir/POETTechnologies/forums/discussion/topics/738135-transcript-from-ld-micro-virtual-conference/messages/2261194#message