Good post but I believe we are talking about specialized equipment that is readily available but may not be part of the fabs existing line or that POET wants under lock and key. In addition I would expect that POET would need their own automated pick and place tools and automated optical testing bench. High volume deposition and etching would already be in place in the existing lines.

As has been discussed previously it is a backend process where the III-V active devices are fabricated separately (at DL) or other facilities. Placed on the dielectric interposer then the dielectric passive elements (waveguides) are processed on the optical side and metalized on the electrical side. This is really reason why yield can be much higher than existing processes. They know both optical qualities and electrical qualities of the dies before they go through the backend process. 

There are still big gaps in understanding but there is enough there to give the overview necessary for investors to understand the value proposition. It is truly a CMOS compatible platform unlike the other platforms that exist. It is interesting to view just how tiny this optical engine is when compared to existing technologies that it is expected to replace. It really equates to many thousands of optical engines per 8 inch wafer.

1mm X 3mm 

To be clear one of the fundamental differentiating factors is that this process is silicon CMOS compatible which enables high volume and low cost:

Other waveguide technologies [Glass, Silica, SiN] don’t have the capability to INTEGRATE with Si CMOS