POET Technologies Inc. (PTK) Earnings Call Transcript & Summary

Our next presenter is a design and development company, offering integration solutions based on the POET Optical Interposer, a novel platform that allows the seamless integration of electronic and photonic devices into a single multichip module using advanced wafer-level semiconductor manufacturing techniques and its packaging methods. It's my pleasure to introduce POET Technologies.

Good morning. My name is Thomas Mika and I'm the Executive Vice President and Chief Financial Officer of POET Technologies. Thanks for attending this presentation, especially to those of you who are on the West Coast as I am. To our current shareholders listening in, this will be somewhat familiar to them. At our Annual Meeting last week, Suresh Venkatesan, our CEO; and Vivek Rajgarhia, our President and General Manager, gave an excellent presentation. My presentation today is directed at investors who are not already POET shareholders. My hope is that in the short 20 minutes we have, I can convey some of the clarity of vision and mission that was communicated last week. We're then at an important moment in the company's history. Let me take a moment to tell you about Suresh and Vivek. Suresh came to POET in late 2015 after 18 years as Chief Technology Officer of GLOBALFOUNDRIES, one of the largest semiconductor companies in the world. He is an expert on all things semiconductor. Vivek joined POET late last year from MACOM and previously was with many of the most well-known optical device companies, bringing many products to market. This is really an extraordinary combination of both technical and management talent. All the risks associated with investing in our stock are outlined in detail in our filings on SEDAR in Canada and on sec.gov in the United States. I'm going to cover some of the highlights of the company, the markets we serve, the challenges that our solutions address and the timing and road map for our products. If there's time left, I'll be pleased to take any questions. POET is a design and development company in the area of photonics. We are rich in intellectual property and talent. We have 69 patents and several applications in the works. We're headquartered in Toronto, so we trade on the TSX Venture Exchange and in the United States on the OTCQX. Our operations are located in Allentown, Pennsylvania and Singapore with another about opening in Shenzhen, China. We chose these locations specifically because they are distinguished as centers of excellent and pools of talent in both optics and photonics. For the past few years, POET has basically been pre-revenue, engaged in developing a unique and disruptive platform in photonics. But to build a company, it's not sufficient to have great technology. The company needs products and customers in the areas in which we compete, the companies are large. So we must be able to demonstrate the ability to deliver product and quantities that are meaningful to them. In our case, that's in the hundreds of thousands and millions of devices annually. Recently, we announced that we had signed a letter of intent with Sanan IC, subsidiary of Sanan Optoelectronics, the world's largest producer of compound semiconductor wafers and producer of LEDs and lasers. In a deal valued at USD 50 million, we are able to achieve both a vertical integration and a consolidation of our assembly and test supply chain. With the size and reputation of Sanan, no potential customers question our ability to scale rapidly and to supply products in large volumes. JV with Sanan is also a direct endorsement of POET and confirmation of viability of our foundational product. We expect to sign a definitive agreement with Sanan IC this month. We serve very large markets where the trends are all positive for rapid growth. These are markets to which we bring a distinct and sustainable competitive advantage. Data center operators, such as Alibaba and Tencent, Facebook and Microsoft, have committed to vast investments in expanding data centers. Neuromorphic optical computing systems are more efficient at supporting inference engines and artificial intelligence than conventional computing platforms. 5G telecom will push computing to edge applications for machine-to-machine communication and the Internet of Things. Each one of these markets will drive the use and density of optical fiber in place of copper because it's cheaper, faster and better. And where there is optical fiber, there is a need for photonic business and a corresponding need for what POET can offer. But photonics applications require the integration of multiple devices that do very different things and are manufactured using very different processes. The different devices include electronic devices such as controllers and amplifiers and ASICs, photonic devices in at least 2 varieties -- we call them active devices, there are lasers and detectors that are powered and passive devices that are not powered such as waveguides and multiplexers. Finally, certain types of photonic devices also need conventional optics to focus or reflect light such as lenses and mirrors. A long-standing challenge in the photonics industry is how to combine and integrate these multiple and diverse devices at an affordable cost and in high volumes. The traditional method of assembling them has been one at a time, using materials that absorb much of the light passing through. This means that connecting the light going between one device and another has to be optimized with each placement. So it's not to lose even more of the light signal. Optimization following each placement requires vast arrays of test equipment and people. Only recently have the largest of photonics device producers been successful at reducing some aspects of the assembly process to the scale of wafers that are produced in the same manner as normal semiconductors have been for a half century. However, even those efforts are still not fully wafer scale. Enter the POET Optical Interposer platform, approach to integration and unification of all 3 categories of devices, assembled and tested completely at wafer scale, the use of a novel waveguide material and a unique inter-design allows a full range of devices to be placed on the Interposer which itself includes a variety of passive devices built into the waveguide material. Different devices communicate with each other using high-speed copper and aluminum traces built into the lower layers of the silicon wafer on which the Interposer is fabricated. With a waveguide material that absorbs very little light and with alignment features built into the platform, alignment of components is assured, and there's no need to test after each placement. The ability to assemble this device at wafer scale, meaning at hundreds at a time per wafer is foundational to the cost and competitive advantage of the POET Optical Interposer. I'd like to review a few the unique features of the Optical Interposer that make it so disruptive. While you may not recognize the feature, you should recognize the benefit. I've already mentioned complete wafer scale assembly and test. The benefit is a dramatic and sustainably lower cost than any application that uses an Optical Interposer compared to a conventional assembly. Chip scale packaging means that we have, by far, the smallest form factor for optical engines. Size is important in electronics, and in this case, smaller is better. Submicron component placement accuracy and low loss waveguides add up to the ability to eliminate the costly testing procedure after each device placement. The material that we use for waveguides and other passive devices is athermal, meaning it doesn't drift when it heats up. By drift, I mean that the frequencies of the light don't shift as the material heats up, which is typical in photonic devices. That eliminates the need for additional temperature control devices, lowering the cost even further. Embedded copper and aluminum layers allow high communication among devices. The benefit is that more of the electronic devices can be included on the Interposer platform, adding more value to it. In fact, we believe that beginning with 400G and possibly 800G in data centers, this will be a requirement as traditional wire bonding will be inadequate to handle those speed requirements. This, again, adds more value to the Interposer platform. Finally, CMOS compatibility means the proposed Interposer can be built entirely with standard semiconductor manufacturing processes which is not true for many photonic devices today, inhibiting full integration and wafer-scale processing. At the beginning, POET chose the strategy of developing a platform technology rather than their product. Recognize the benefits of the materials and design, we understood how flexible and adaptable it could be. So instead of engineering something to be used for 1 product, like a one-off 100G CWDM4 Optical engine for a transceiver, we thought about many potential applications for the Interposer and developed a platform technology instead. It took longer, but we now have the benefit of being able to easily adapt the built-in features of the Interposer to a range of products including generations of the same product, product extensions, derivatives and enhancements. This avoids having to reengineer a product from the ground at each time, and we continue to add features that make it more adaptable to other applications like the ability to add lenses and collimators at wafer scale to the platform, features that will be required for 5G applications. An example of a platform approaches are designed for an optical engine for transceivers that scales from 100G to 200G and then to 400G without redesign. All that is required is replacement of the components. The platform itself functions in the same way it did in the previous generation without the need to invest again from ground up. Just to review, a transceiver is a device that converts electronic signals typically carried on copper, emanating from a computer chip, for example, into light signals and back again at both ends of an optical fiber. And optical engine is the value-added portion of a transceiver that deals with the meeting light, modulating it, guiding it and detecting it and converting those signals from electronic to photonics and from photonic back to electronic. Here are the actual completed Interposer designs for the products that we are now developing in connection with proposals and engagements with customers. Multiplex light engines are for use with silicon photonics assemblies. Shown here are 2 varieties, 1 using our annular waveguides and another using a different type of demultiplexing filter. We already discussed the scalability of the optical engine for 100G, 200G and 400G, which appears in the center. Finally, we have a receive optical engine for 400G that some customers are interested to replace existing modules, subassemblies within their 400G transceiver modules. We have thought hard about how to protect our fundamental IP while being able to produce devices at scale. So our manufacturing strategy reflects the belief and the versatility of the POET Optical Interposer. Interposer IP is oscillated with a small team at POET and is manufactured by our foundry partner, SilTerra Malaysia, where we own the critical equipment and the recipes. We will supply completed devices to the joint venture or to whomever is assembling that final optical engine. We source photonic devices from -- strategically from select partners, including Almae in France, DenseLight in Singapore, our former subsidiary, and Sanan in China. Most of these devices are POET designs. We have engineered the platforms that we can use off-the-shelf or customer-specific electronics. And with the JV, we consolidate the current assembly and test supply chain into a single entity that is just under 50% owned by POET. To give you a sense of where we are in our company life cycle, during 2020, we have fully transitioned from technology development to product development. In this transition, we move proving feasibility for the technology building blocks associated with the Interposer, which we did in connection with a large North American customer to designing products to specific customer specifications, and that's the stage we're in now. This results in a much more deterministic schedule for product development than we could put forward previously while we were in technology development phase of the company. Each of these 4 products has at least 1 and, in some cases, more than 1 customer engagement that is driving the product development. As we move through alpha stage in which we provide samples to customers to beta where the samples meet customer specs, we expect engagement to turn into contract. This is happening now, and it's accelerating both in terms of the pace of progress and the interest of customers. Today, POET has the team, the technology, product designs and the customers that are committed. As we get into the early months of next year, the ability to manufacture in volume will become critical. We expect to have the JV agreement completed soon so that we can get facilities and the team up and running to supply initial orders by mid-year. Over the next months, we plan to make regular announcements of progress against our plan and several of those announcements will include specific customer engagements and contracts. Thanks for your time. And if there's any time left, I can take some questions.

There, in fact, is, Thomas. We've got quite a few. We'll hopefully, try to get through some of them here today. Can you update us on the tender progress with the North American Tier 1 customer?

That process is in -- is still in the stage of request for information. It has not yet been issued as a formal request for quotation. We think that is a result of the fact that the specifications are -- seem to be changing over time.

Yes, that makes sense. The sign joint venture with Sanan, is it still on target for the end of this month?

It is. Yes. We're working diligently towards that. And we have every expectation that the joint venture agreement, at least, and the licensing agreements can be signed.

Good. So do you want to talk about the competitive landscape, right? Are the competitors operating in the same space as POET, maybe talk a little bit about that, if you could?

Well, one of the interesting things about what we're doing is that the transceiver module suppliers, some of them are long known in a light momentum AOI, are going to continue as suppliers to their existing customers that are systems companies and in some cases, data centers. What we're supplying is a component. So what we're not going to be doing is building a full module and competing head to head with some of those much larger companies. Instead, we're offering them a component at a much lower cost, higher performance. And we believe that we will -- it'd be assisting those companies that adopt our component versus whomever else they're procuring multiple components from that it will essentially extend their business model and their product license. So that's a very different approach than say, for example, competing directly with a module manufacturer.

Yes. Good. Yes, that makes a lot of sense. a fairly robust pipeline of products. Can you approximate the number and types of additional products that may be added in the pipeline over the next 12 months?

Well, we're actively engaged with another company for the development of a platform in the artificial intelligence arena. And that looks an awful lot like what's called co-packaged optics, which is bringing the photonics devices as close as possible to the computational devices. And the reason that's happening, particularly in network server blades and so on that are in data centers, that the speeds of the switches are increasing so much that dealing with the signals coming out of those switches is consuming the energy budgets that are available for those switches. So the architecture of the switches that is the peripheral devices outside of the actual switch itself, which in the latest generation is operating at the 25.6 terabits per second, I think. It's an amazingly fast switch, has to be reengineered. So the architecture of that has to be reengineered to take out some of the intervening devices between the switch signal coming out of the chip and turning it into a light signal. And much of what we're looking at in terms of the platform for these computational engines in artificial intelligence can look very much like what we would have to do for this area of co-packaged optics, which has been under studied by the largest of the data center operators for a number of years in various consortia which we've recently begun participating in.

Perfect. Maybe one last question, if we could you think about capital markets. Do you have any plans to uplift on Nasdaq or another national exchange here?

Not yet. I think that a prerequisite to doing anything like that is having a reasonably predictable revenue stream. And it's at that point that we would consider that kind of move. That has to be done in a very positive and careful way if we were to do it all.

Yes. No, that's prudent. Maybe just a minute left, I'll turn it back over to you, Thomas, let you just have some closing remarks for folks today, and let everybody move on to one-on-ones in the next presentation.

Yes. so thanks very much. I know that this was partially pretty technical for some people. But at some point, when you're a technology company, you eventually have to resort to a technical description. So if there's anybody out there that has any additional questions, you can go to our website, which is quite a good explanation, and you can just send me an e-mail because I'm happy to answer any such questions. So thanks very much for joining today, and I look forward to talking with you again soon.