Just remember that many will need to define their finished product based on intended use and environment. Some may want a relaxed standard accepting a lower max operating temperature based on the specific data center design (ie Facebook) in an effort to reach the 1$/G target.

What they have probably already received is the engineering samples of the alpha design which provides them with a basic prototype (not the finished product to meet their data center design needs). And that is important because it provides the baseline for both the physical metrics such as insertion loss and the ability of the dielectric waveguide to collimate (trap) the light in the specific wavelength band required across the temperature range they are specifying for each application. And of course they will need certainty that POET’s capacity to get the job done is in place (and can be expanded if required without significant ramp delays). Remember that POETs platform process allows the assembly of known good die so yield is enabled to a great degree by the steps which mean a lower production cost. Along with the removal of free space optics and that manual intervention means the lower production cost benefits everyone and becomes a main driver. 

I read oz's great post re Accelink update several times and they have some very interesting projections.

I think we have the right bait, the right spot and the right time to get few trophies.

https://agoracom.com/ir/POETTechnologies/forums/discussion/topics/714277-accelink-update/messages/2205728#message