I think the beauty in what Suresh claims, he's actually doing it at wafer scale, cheaper and smaller form factor, I'm sure the fact you need 10x more volume for the AI process makes the form factor and cost very important. When the investor group handed Celestial AI the 56 million cash, I'm sure they got a peak at the light engine (and the whole AI Accelerator) and really liked what they saw.
It's not just a concept founded in a university lab, these solutions are all prototyped and we can tell by the hiring lately, they're going into production this year as there's going to be a lot of interest for customized solutions like LR4 FR4 and these light engines which may be very difficult for the competition (Suresh said the competition will be very little at the 2021 AGM in LR4 and the same prospects for upcoming FR4) but as you just heard Suresh now, it's easy for Poet because the platform is ready to go.
SP