So I ponder sometimes, how could the OI be improved, and POET get sideswiped and become the Beta vs. VHS, or just miss the boat vs. some magical tech. 

I believe Foxconn, CAI, Mitsubishi, Luxshare, MultiLane and the rest have searched every corner of the earth and concluded that POET is the answer. 

Think about it. Could it be smaller? Not much, if at all, then the modulaters, lasers and other specialty items would be more challenged to find a solder bump or other mounting spot. The OI is half the size of your little fingernail, so I think it is right sized.

Could it have better waveguides? I was told that it beats silicon by 10X in this metric. They have a patent for however they are doing it. An incremental improvement will not beat POET. 

Ditto for insertion losses (very low), fiducials built in, flip chipping, wafer scale testing, out of plane mirrors and maybe some things I've missed or are secret. 

All this at prices that are nice for us but break the backs of the competition

The point is, it is as good as necessary, and with mass production and multiple design wins, it becomes a space to be ceded and not worth trying to win it. Is anyone beati.ng ASML oon their specialty tech. Nope. Is TSMC getting hammered? Nope. NVIDIA is getting all the glory, and the competition is fierce, but way behind. 

Is POET like the above? Kinda is. Does it really need to be? Once the biggest players and most sophisticated mid-level (think MuliLane, Adva) companies, plus the newcomers that are CAI or the like, have cast their votes, what is there to do, but get in the queue and design in.

That is where I think we are headed, and the orders will reflect when they come in volume.