What it looks like to most of the semiconductor world is that 2018 is "peak revenue" year for Fab8. It will never, ever generate as much revenue as it does this year. With no new technology on the horizon, and the key 14LPP customer moving to TSMC 7nm, the ASPs and wafer numbers will decrease from this point forward. You can cry "derivatives" but what is even defined yet? Isn't the general consensus that Fab8 needs to capture something like 25-30% of the derivatives over the next few years even to keep the doors open. That's a big order and the whole idea of the differentiated products is more products from more customers and the management at Fab8 has no idea how to handle that, the main focus has been on 2 1/2 customers (AMD, Qualcomm and IBM 14HP which is 1/2 done in Malta).
My hope is something is able to work out for all my colleagues that are still there, I just don't see it, the numbers don't work. The company's already set itself up to show that if something is too hard or doesn't work out quick enough, they will abandon the project. So who's going to sign on for any of the early derivative flows given Fab8's history of throwing the towel? AMD has gotten screwed twice with 20nm and 7nm, putting in resources, product designs, etc only for the project to be abandoned. The best outcome is that some other company buys Fab8 and it is repurposed (unless it is Samsung and they can keep running 14LPP).
Reposted this from @WdD53bY-5grj for being on point