As far as I understand it. They're giving up on the 4 terminal Thyristor (power switch?) in their integrated POET devices. They got 2 & 3 terminal versions working well, but couldn't make the manufacturing work for the 4 terminal version. They've already started on a "workaround" of a bipolar based switch design, but it seems like it still needs to be ironed out (proof-of concept is good) and then the whole shebang needs to be re-optimized to work with the new design. The part that freaks me out a bit is that they seem to have diverted almost all resources away from the POET portion in the hopes that someone will step in and partner up. In between the lines, it seems to me that the development cycles on the GAAS stuff are too lengthy and costly until they have someone with in-house GAAS fab capibilities onboard.
IM (non-expert)O