BUSINESS - Overview - We design, develop, manufacture and sell both discrete and integrated opto-electronic solutions for the sensing, data communications and telecommunications markets. In addition to manufacturing a range of Indium Phosphide (InP)-based light sources, POET has developed and is marketing its proprietary POET Optical Interposer? platform. The POET Optical Interposer utilizes a novel dielectric waveguide technology that allows the integration of electronic and photonic devices into a single multi-chip module. The integration of devices into a single package is achieved by applying advanced wafer-level semiconductor manufacturing techniques and novel packaging methods developed by POET. POET’s “photonics in a package” eliminates costly components, assembly and testing methods employed in conventional photonics solutions. In addition to lowering costs compared to conventional devices, POET’s Optical Interposer provides a flexible and scalable platform for a variety of photonics applications ranging from data centers to consumer products.
POET’s Optical Interposer is a platform technology upon which multiple applications can be based, including transceivers for data- and tele-communications, integrated photonics on electronic switching devices, low-cost components for the networking and cellular markets, automotive LIDAR and a plethora of sensing and other applications using light as a medium for data transmission. In each case, devices traditionally associated with photonics, such as laser diodes, light emitting diodes, detectors, amplifiers and the associated waveguides and other passive devices are designed specifically in the context of the Optical Interposer to meet the needs and functions of specific applications.
POET has targeted as the first application of the Optical Interposer the development of Optical Engines for transceivers. Transceivers are used to convert digital electronic signals into light signals and to transmit and receive those light signals via fiber optic cables within datacenters and between datacenters and metropolitan centers in a vast data and tele-communications network. We expect to deliver prototypes of certain types of Optical Engines by mid-2019. These prototypes are expected to address a small portion of the market for Optical Engines. Continued development of the prototypes will add several commonly used communication protocols and data speeds, to increase the functionality of the Optical Engine from receive only to transmit and receive, thus improving device capability and performance. Concurrently, we also intend to develop additional applications for the Optical Interposer platform in the telecom, LIDAR and sensing markets.
Research & Development - Virtually all of POET's R&D expenditures in recent years are in some way connected to the Optical Interposer. We expect to continue to spend the large majority of our R&D resources for the foreseeable future on Optical Interposer-based products across a wide variety of potential applications. The only other R&D expenditures that we have or may incur relate to conventional non-interposer-based products that we develop and manufacture for our legacy sensing product lines that represent the majority of our current sales. However, we intend to develop and transition these products to the Optical Interposer, because of the resulting cost and performance advantages that it provides.
POET’s Optical Interposer development program consists of over 20 development projects in three areas: 1) Active Component Development, which includes a variety of application-specific Indium Phosphide (InP)-based lasers, detectors and modulators; 2) Passive Component Development, which includes application-specific filters, mux-demux devices, waveguides and spot size converters, all designed and fabricated using POET’s proprietary 3 dielectric materials and processes; and 3) Core Integration Process Development, which includes processes such as assembly, hermetic sealing, flip-chip techniques, reflection management, and wafer-level test. In order to optimize our development resources, we have taken a “building block” approach, beginning with the most fundamental functions needed for the Optical Interposer in each of these three areas. The Optical Interposer is unique in the industry, incorporating several “first time ever” implementations of advanced optics and semiconductor packaging techniques and completely new, novel designs for components. To minimize risk and maximize the probability of successful outcomes, we run parallel development programs, both internal and external. Our external programs engage development partners or subcontractors to provide devices, process expertise or equipment that we do not have internally.
As a platform technology, Optical Interposer development does not have a specific end-point. Each application of the Optical Interposer requires development specific to the application. POET’s product roadmap is currently focused on the development of Optical Engines for optical transceivers. Optical Engines include all of the photonics-related components of a transceiver but does not include several of the electronic devices for a functioning transceiver module nor does it include the external packaging and optical fibers. The electronics include such devices as Trans Inductance Amplifiers (TIA’s), laser drivers, etc. that are produced by major semiconductor manufacturers. However, Optical Engines represent the majority of the cost and value of most optical transceivers.
The “active” components that are included in a POET Optical Engine include lasers, detectors and modulators fabricated on InP substrates. To exploit the unique functionality of the Optical Interposer, each of these devices must be made to a design that integrates spot size converters (“SSCs”) and allows the device to be compatible with a flip-chip assembly process. Our DenseLight subsidiary has been engaged for the past two years in the development of designs and process technologies to build such devices for the Optical Interposer. To accelerate the development process, we have either combined efforts with development partners, purchased wafers to specific designs or licensed technology as a means to supplement our internal development efforts. One of our earliest internal developments is a QuadPIN photodetector, which was introduced to customers for qualification in late 2018 at 25G1 speeds. We believe the performance of this device can be improved to allow it to be used at 50Gbps speeds, which would be used in 400G transceiver Optical Engines. As we continue that development, we intend to place this device on an Optical Interposer with an integrated thru waveguide in Q2 of 2019, primarily as a means to demonstrate the functionality and versatility of the Optical Interposer platform, rather than as a product for sale. We have supplemented our active component device development with co-development partners and license agreements, including for certain types of lasers and modulators. In particular, we have initiated the development of Optical Interposer-compatible components at Almae, one of our laser development partners. This not only mitigates the risk to internal development and accelerates time to market, but it also ensures a second source of Optical Interposer-compatible active components, a critical part of our strategy going forward.
In parallel to these activities, POET has also directed development programs in the other two areas outside of DenseLight, including Passive Component design and development and Core Integration process development. Passive devices, as mentioned above, include filters, mux-demux devices, waveguides and spot size converters, all designed and fabricated using POET’s proprietary dielectric materials and processes. We recently established a waveguide development lab in Ottawa in association with Mill View Photonics. We plan to expand that effort over time in order to design waveguides for specific applications for the Optical Interposer across several vertical markets. The actual fabrication of the passive devices, which are built on 8-inch diameter silicon wafers is performed by our foundry partner, SilTerra Malaysia (“SilTerra”). The devices fabricated at SilTerra represent the base foundational elements of the Optical Interposer on which the active devices are placed. In early 2018 we transferred the basic processes for the deposition and patterning of our proprietary dielectric material from a university lab to SilTerra. We purchased dedicated equipment in order to preserve the intrinsic intellectual property of the processes, and since early 2018 we have continued to improve those processes in order to make them suitable for high volume manufacturing.
1 “G” is an abbreviation for “Giga bits per second”, the rate at which the device transmits or receives data.
The third area, Core Integration Process Development, highlights the fundamental benefits of the Optical Interposer platform as primarily an advanced packaging technology that allows true wafer-scale assembly and test. We do not believe that such true wafer-scale integration has been achieved by any other technology in the photonics industry. We are able to achieve chip-level integration and wafer-scale assembly, test and packaging because all of the active devices include SSCs and are designed to be placed and “matched” to passive device interfaces on the foundational Optical Interposer wafer using pick-and-place assembly techniques. We achieve high levels of coupling efficiency between each device, eliminating the high cost and cumbersome process of testing each component following placement. Once placed and tested at wafer scale, each Optical Interposer device is sealed, the wafer is separated into hundreds of individual die and the final Optical Engine is ready for shipment to the customer. Each of these process steps, from flip-chipping of devices onto the dielectric, pick and place assembly, hermetic sealing and singulation required substantial innovation and development, including several techniques that are unique in the photonics industry.
In late in 2018 we were approached by two large global networking companies that saw promise in using all or part of our Optical Interposer technology for their 400G transceiver development projects. These projects offered the benefits of payment for development, intersection with major companies at the early stage of their development projects for next generation transceivers, the prestige of working with industry-leading companies, and the potential for our Optical Engines to be included as major components of the planned shipments of transceivers by these companies. We took a major decision to reorder our priorities in order to effectively support these companies. We believe that addressing product-specific requirements with willing partners and committed funding is the optimal way to introduce the Optical Interposer technology to the market. Identifying and overcoming individual technical challenges increases the likelihood of success and promotes innovation. We expect that successful implementation of our designs into component prototypes, combined with the cost advantage inherent in Optical Interposer-based solutions, will lead to additional funding for other products, as well as to contracts for the delivery of production devices, once fully qualified.
The immediate consequence of our decision to work with leading industry partners on 400G transceivers was to accelerate the development of higher performance lasers, modulators and detectors needed for 400G, at the expense of similar programs for 100G. Another factor in our decision-making process was the market forecast for 100G transceivers which has flattened in terms of revenue and appears to be maturing much faster than the industry anticipated. Industry total revenue for 100G transceivers in 2019 is expected to be flat or lower than 2018. With unit volume going up, pricing is down and therefore margins are squeezed even more heavily than before.
Our plan to deliver 400G devices is essentially unchanged from prior plans, which targeted the release of TROE 5 Optical Engine prototypes to customers for qualification2 for in the second half of 2019. Our revised plan calls for all of the required active components, waveguides for certain standards, and core integration processes to be far enough along to allow the Company to produce both 400G prototypes and 100G prototypes during the second half of 2019. We expect to be able to introduce the less complex standards, such as DR4 for 400G and Parallel Single Mode 4-fiber (“PSM4”) for 100G earlier in the second half than the more complex FR4 for 400G and Coarse Wavelength Division Multiplexing (“CWDM”) standards. In each case we have multiple parallel programs aimed at these prototype products, utilizing both internal and external development resources.
Further, while we had forecasted the completion of a Receive Only Engine (ROE) for 100G as our first planned prototype, we now expect that there will be little demand for a separate ROE, since customers will be able to qualify a full TROE instead, performing one rather than two qualification cycles for a more complete solution. However, a 100G TROE with a CWDM filter represents the second generation of transceivers being adopted and a standard which is expected to be dominant in the 100G datacom market in China. Early adopters of 100G transceivers in the United States utilized PSM4, a standard that does not incorporate the highly complex CWDM filter.
The inherent cost advantage of an Optical Interposer-based 100G Optical Engines should still allow entry into the 100G market, but the level of penetration will depend on how far pricing will have fallen by year-end and into 2020. Nevertheless, because of the size of the market and the need for transceiver module suppliers to address margin concerns, an offering of a 100G CWDM TROE could still have major impact on POET’s datacommunications revenue over the next one to two years. Due to the relatively large amount of NRE in our 2019 forecast, we do not expect to have to revise our guidance for 2019. Overall, we believe that our revised development roadmap represents a sound foundation for growth in 2020 and beyond.
In addition to the new products in the data communications market noted above, POET expects a significant increase in its Sensing product revenue in 2019. Recent trends have been positive. With some existing customers expanding order size and the addition of new customers, we have a strong pipeline of committed orders. Expanded laser product-lines in CW, FP and DFB configurations and in high-value NLW modules which address new and complex applications in the sensing market are in high demand. In addition, by incorporating the packaging innovations developed for the Optical Interposer, we expect that in mid- to late 2019 the Company will be in a position to introduce a differentiated sensing product-line at lower cost and higher performance.
In recent quarters, POET has taken major steps to advance its development of Interposer-based new products, including through the purchase of equipment, improvement of facilities and the strengthening of its engineering team with more highly qualified talent and larger staff, all represented in POET’s consolidated financial statements through additions to fixed assets and increased operating expenses. Certain additional capital equipment may be needed to enhance our development and production capabilities, but we expect only marginal increases in operating expenses over the next 3 to 4 quarters, as we are able to address the needs of our customers with our existing engineering staff and production facilities. The next few quarters will be devoted to the successful completion of funded development programs, the introduction of new devices into qualification cycles with customers, and preparation for higher production volumes in subsequent quarters. As a result of the recently announced contractual commitments for Optical Interposer sub-assembly prototypes, other anticipated Optical Engine prototype orders in late 2019, increases in Sensing product revenue and improvement of overall gross margins described below, we expect our DenseLight subsidiary to have a material increase in revenues in 2019. Should the Company’s structure continue with DenseLight as a subsidiary (see “Potential Sale of DenseLight Subsidiary” below), forecasted revenue would be in the range of approximately $8 to $10 million, though we do not expect to reach cash break-even until 2020. Gross Margins should increase as a result of the proportion of higher margin development contracts for Non-Recurring Engineering (NRE) compared to products sales, which utilizes POET’s existing engineering and operations staff. In addition, we plan to continue product development with an expansion of opportunities in markets beyond data communications for our Optical Interposer technology, such as telecommunications, Automotive LIDAR, and integration with Application Specific Integrated Circuits (ASICs), including switches and graphics generators.
2 “Qualification” of new devices or components, demonstrating adherence to both customer specifications and industry standards, is done both by POET and by our individual customers. The period over which device testing occurs may extend from three to six months or longer, depending on the device, the qualification tests required and the customer. We have estimated that the qualification of its “passive” devices, which do not contain “active”, light-emitting components should average approximately three months and that active device-containing components should average approximately six months. However, with any particular device, test or customer, the qualification period may be shorter or longer.
