From the Zack's reearch article:

POET has been trying to show customers the radical improvements its platform can offer them and will be providing demos with its partner Siluxtek at its labs located near the China International Optoelectronics (CIOE) trade show. This show is being held September 16-18 in Shenzhen. It is the world’s leading optoelectronic exhibition and more than 3,200 exhibitors will present the entire optoelectronic ecosystem there including information and communication, precision optics, sensing, lasers, infrared and photonics.

At this show the two will demo the 400G FR4 transmit optical engine and samples will be available beginning in Q1 2022. Its 400G receive optical engine will also be demonstrated at the show. While the 100G CWDM products sold into China will generate the first revenues for the joint venture, the 400G FR4 product is expected to provide the first revenues direct to POET possibly by Q3 2022. 

While POET’s 100G CWDM4 optical engine will be the first product to ship it will be the lower margin, higher volume product for the company as a transceiver module sells for about $150 in that market. POET’s solution could bring the transceiver price down to near $100 and might generate $50 - $75 per unit sold. However its LR4 engines (long range) are going into product that costs three times more and POET’s ASP will he much higher and margins will be much higher. Plus more than one of POET’s 100G LR4 engines can go in a module. A single 100G LR4 transceiver costs as much as $350, with POET’s pricing for its optical engine being about $150 per unit in low volumes. With two 100G optical engines that could be $300 and POET believes it can engineer up to 4 engines in a module for $500-$600. At a 50% margin is could make a gross margin of $250-$300 per unit compared a $25-$30 per unit gross margin of the 100G CWDM4

BTW, where's BCD?