Because most people who use Python use it not because of the language itself but because of its library / framework eco-system and PyPy's compatibility with that is limited.<p>Also PyPy offers no performance advantage if all the CPU intense stuff happens in libraries written in C/FORTRAN/ASM anyway.
In many cases people don't care about performance (CPU time) but they do care about problems, real or imagined, if they try something new. (Human time)