A point missing from this line of discussion goes to the very root of why ISRO woud be interested in this POET's technology in the first place.  This isn't just about the Government of India acquiring a game changing technology in the event it could be made to work, but more specifically an organisation looking for a solution well suited to their needs.  POET's GaAs offers this to ISRO.

Due to its inherent properties GaAs is a substrate of choice for space applications and the broader aerospace industries, coupled with the other advantages Taylor's platform brings, IMO, there is plenty of incentive for both to make efforts to develop the technology.  In this context, the GaAs market is more niche than InP (ie low volume, high value) but still very important to these industries, and applications particularly suited to GaAs in space or aerospace will not be replaced by InP based solutions anytime soon.  In a general sense, the GaAs market is developing with significant growth expected over the next 5-10 years so makes good business sense to continue efforts on this front.

As FJ has indicated, in addition to the GaAs tender itself, we have seen related tenders (ie GaAs wafers for planar optoelectronic technology, etc) which suggest a much more comprehensive process going on, and likely ISRO is (or was?) very interested in the technology.  IMO, I believe there is more evidence supporting the GaAs tender vs against at this point.