I prefer this sentence:

The 800G Rx has been well received, fully qualified and is expected to be incorporated in the optical transceiver modules of several customers this year. 

POET targeted as the first application of the Optical Interposer the development of optical engines for optical transceivers used in internet-based data centers. Optical Engines include all the passive and active components related to the production, manipulation, and detection of light within an Optical Transceiver. Optical Transceivers plug into switches and servers within the data center and allow these network devices to send and receive data over fiber-optic cables. We chose this market because it is large in size, has established standards for device performance, and the unit volumes of devices shipped annually are exceptionally high. It is a market in which our advantages of cost, power consumption and ability to scale rapidly allow us to be competitive with other suppliers.

The rapid growth of AI software systems represents a profound opportunity for POET. We believe that the rapid growth of software services can only be sustained with hardware that meets the challenges of increasing speed and bandwidth, lower power consumption, lower cost, and the ability to scale to the volumes that will be required by data centers globally. POET meets these challenges in two ways: first, by providing to the market integrated, chip-scale Optical Engines that perform at the levels that are now being deployed in the most advanced AI clusters at speeds of 800Gbs (gigabits per second); and second, by offering what we believe is currently the only viable path to increasing the speeds and bandwidth of Optical Transceivers to 1.6Tbs (terabits per second) and 3.2Tbs in industry-standard pluggable form factors. In addition, we have used our Optical Interposer technology to develop Light Source products that address newly emerging architectures in data centers that are based on chip-to-chip data transfer using light, rather than electrons, which resolves speed, bandwidth, heat-generation and cost issues at a fundamental level. The combination of POET’s focus on leading-edge Optical Transceivers and Light Source products for next generation data center architectures essentially places POET among a small number of suppliers globally that are truly “pure play” AI hardware companies.

Research & Development

Beginning in 2017, POET began designing lasers for data communications applications and directed DenseLight Semiconductors, Pte. Ltd., a former subsidiary of the Company, to build such lasers to be compatible with the Optical Interposer platform. In 2019, the Company decided to adopt a “fab light” strategy, common among semiconductor companies, and divested its fabrication operations through the sale of DenseLight in November of that year. From 2018 - 2020, virtually all the R&D spending in the Company was dedicated to design & development of the Optical Interposer as a versatile platform technology, replete with features that enhance its utility across a variety of application spaces.

During the second half of 2021, the Company transitioned to product development by investing more than $2 million in the design & development of 100G and 200G optical engines in several configurations, including customized designs for specific customers and applications. Samples of optical engines at various stages of development were made available and delivered to customers in 2022 for initial evaluation and in 2023 for design-in and customer qualification. SPX is forecasted to produce Optical Engines in high volumes for several customers in 2024. POET’s effort in lower speed Optical Engine design and production was intended primarily as a way for POET to demonstrate the viability and market acceptance of its unique approach to integration and fabrication and to establish an initial presence in the market. However, the Company’s primary strategy is to offer Optical Engines at the highest speeds at which customers are deploying Optical Transceivers. In 2024, we expect that we will be primarily in 800G, and heavily focused on those hyperscale data centers actively implementing AI services. Consistent with this strategy, the Company has invested approximately $20 million in design, development and engineering programs related to its 400G transmit chiplets (combined in multiples of 400G to achieve 800G, 1.6T and 3.2T speeds), in 800G receive optical engines, and in light source products, and fabrication techniques.

The Company has designed, tested and sampled the current version of its 400G transmit (Tx) engine, and its 800G receive (Rx) engine with various customers. The Company intends to revise its 400G Tx product and to introduce a new version later this year. The 800G Rx has been well received, fully qualified and is expected to be incorporated in the optical transceiver modules of several customers this year. So long as the Company provides Optical Engines to optical transceiver module customers, there will always be customer centric adjustments to these products to fit their specific needs. The cost to make these adjustments will vary depending on the customer requirements.

The Company is expected to invest an additional $11 million in 2024 in ongoing development of the 400G Tx chiplet for inclusion in 800G and 1.6T optical tranceivers. POET is also committed to the development of its own optical transceiver modules, a critical next phase in the Company’s growth plan, with investments in that program beginning this year. At the present time, the Company expects to have a functional module by 2025 with sales of modules ramping in late 2025.