I agree, I have thought the same thing: Why not maximize revenue? Why not offer it for $2000 or something--surely way more people would buy it at that price.
But then I think to myself, perhaps it's not that straightforward. Perhaps by simply maximizing revenue Tesla would at the same time maximize headaches. Maybe Tesla doesn't want everybody being able to afford FSD because it's not ready for prime-time and they know this, they don't want every yahoo buying it and thinking, "Oh gee I got Autopilot, that means I can fall asleep at the wheel and drive home drunk and make out with my girlfriend while filming a Youtube video on my way home after clubbing."
I think anyone looking to spend $10K on FSD would look seriously into what they are purchasing and would conclude one of two things: 1) FSD sucks and isn't worth my hard-earned $10K; or 2) FSD sucks, but I don't care because I'm a Tesla fan and want to see the company succeed, it'll probably work eventually so take my money, and if it doesn't work, oh well, I'm helping out an awesome company.
Tesla knows there are customers like me they can count on to test out a product that simply isn't ready for prime-time, isn't ready for the masses. Tesla already has enough problems as it is, and sometimes stupidly chasing after low-hanging revenue can lead to more problems than it's worth. In any case, once FSD is truly ready for prime time my car's value will jump five-figures easily, and the SP will jump another $100, so it's all good.