The force comes from the compression of the spring, so equilibrium point is when the spring is in it's natural uncompressed state (x=0 is defined as this point).
Sometimes you can define a new equilibrium point for a spring that's got a mass hanging from it. This helps simplify vertical oscillation problems because the spring still oscillates at the same frequency it would if it were horizontally oriented, but it oscillates about its "new" equilibrium rather than its uncompressed equilibrium.