ASML's monopoly on EUV tools, and their high cost, long lead times and low number of units shipped per year are probably the most significant limitation for leading-edge logic fabrication nodes. But even once you have one or two NXE:3400 tools, you still need at least a few years of R&D before you can get to commercially viable manufacturing of sub-7nm chips. (For example, TSMC developed their own pellicles for EUV rather than relying on ASML for that component.)

I think it's just kinda impossible to start one, it's prohibitively expensive to get the factories and equipments running, and there aren't enough people with the know-how of building and operating these things in the world. Not sure if I understood your question, but I hope so