@Quan,

I agree with you. My gut feeling about PET is that we'll need validation via market penetration with VCSELs before we can sell PET to industry. I think using POET for general computing isn't going to be cost effective, what with GaAs being a more expensive material. Unless we are integrating the functions of several chips into one, we can't expect adoption as a straight chip-for-chip swap to replace current Si-CMOS technology.

Of course, there are savings to be had by running POET as a power saver if you and end-user. There also may be customers that want to run PET for superior performance - power savings be damned. There's resistance to radiation for space applications, and PET has POET extensibility over the longer term when we ultimately get total POET convergence. But maybe it's a bit too high a risk with no current product and a long development horizon to sell the idea.

Best we lead with POET VCSELs.