First off most people here have no idea what your reference is canadiancamper.

The question you are asking is associated with the road maps produced by Michael Lebby and presented at the Epic Online Technology Meeting on PIC Packaging.

The purple bricks represent a “Technology Cost Barrier”. So the quick answer is POET will enable the connectivity performance part of that equation through wafer level packaging and testing. But it will be very much up to industry to produce the associated chips (active elements) at a cost and performance level to meet those targets.

The whole point of POETs business case is to provide high performance connectivity, packaging and testing and do it in an automated fabrication flow at a much lower cost than existing systems. Even the end to end testing between receive and send have been enabled at the wafer level using a sacrificial waveguide that connects them end to end for testing before wafer dicing. And I think this might be where the custom specialized tool comes into  play as a modified probe tester or  it  could  be a flip  chip bonder which allows accurate high speed reflow with the die mating  features. Reflow is achieved across the entire wafer. In one step hundreds of die are connected into their positions and permanently fixed into place.

Back on topic: There are of course mods required by POET to the active optical chips for spot size conversion and flip chip mating features but for the most part the plan is to mix and match off the shelf to the extent possible and this again will drive down costs.  POET drives down costs in a number of ways and they will all contribute to achieve industries goals. Having industry leading coupling efficiency means that lower power lasers can be used. And really we come full circle in how Suresh wanted to educate the market on what the fundamental strengths of POET are. For those thhat recall there are no leaking faucets here.  Photons are precious and POET's optical interposer platform does not waste them and in doing so energy to generate the photons and the heat produced are minimized. Also components are minimized no reflection means isolators are not required. I would also  suggest that PDK's are also  simplified as a result of reduced signal loss and reduced parasitics.