Hi
@CriticalThought
Quite an interesting discussion I must confess. You see, the main problem with FABs is that they are very capital intensive! Just to give you a perspective, a FAB producing chips for industrial automation, automobiles etc cost in the vicinity of ~$2.5bn to establish, whereas it takes ~$5bn to set up a fab dealing with top of the line 10-14nm process. Also the data generated by Taiwanese FABs clearly show that the yearly "profit" is hardly in the range of ~$100mn only. So, unless and until Pakistan doesnt have a huge demand of these micro chips, it makes no financial sense to set up a high end fab. By demand I strictly mean that the local vendors should find Pakistani manufactured chips to be more affordable than the one sourced from singapore, Taiwan FABs. The best for Pakistan would be something like setting up 180nm process that can produce ICs for applications ranging from basic industrial automation, routers and other basic stuff like how India did in early 2000s. Then progressively new processes can be incorporated with political will and lots of dollars.
Other area that can be exploited is the GaAs or GaN FABs that find direct application in AESA radars. But then again the main challenge here is the "demand". Pakistani armed forces should have an adequate demand of AESA radars--all kinds right from LLTRs, MPRs, HPRs, AEWCs to fighter borne radars. Although not as expensive as microprocessor manufacturing, these FABs still cost something like $300-500mn to erect(as India is doing at IISc Bangalore). But FABs are just the one piece of puzzle in AESA game. Modern AESA radars are bottlenecked by their processing capabilities and algorithms than by hardware as modern experience shows. Although not strictly AESA, I have indeed worked on PESA for a short stint as a internee. Once you have mastered the hardware, the real challenge of making an AESA, a proper AESA starts. The research trend is heading towards developing novel algorithms for stuff like adaptive beam forming leverging the capabilities of a digital beam former in modern AESA radars. Concluding my post, I wish to highlight that hardware is just one piece of the puzzle, the main challenge these days lie on the algorithm side!