This may not always be true. Consider a piece of nickel wire at above its Curie temperature, where it is non-magnetic. The electron spins are randomly oriented and there is no net momentum to the electron gas as a whole - just as many electrons are moving in one direction as in any other. Below the Curie temperature the latter is still true, but now the electron gas has adjusted itself to a lower energy state since the spins have now oriented themselves.
Edit: I did some further investigation on this and it seems that this may not be a good example. The ferromagnetic transition is of course second order, so there is no latent heat release during the transition. For iron, it seems that most popular idea for a while was that the electrons reduce their coulomb repulsion at the expense of increased kinetic energy at its ferromagnetic transition. But a paper came out in 2014 that challenges that and concludes it is the opposite.
https://journals.aps.org/prb/abstract/10.1103/PhysRevB.90.125102
So there is some controversy in this basic system! The point is I think, though, that the total energy of the ferromagnetic system is a constant and there is just an exchange, one way or the other, between the potential and kinetic energy of the system. I don't think this point is usually covered at all in most basic textbook discussions of ferromagnetism. At least, I can't find it in any of mine.