How can you prove that the normal equations: $(X^TX)\beta = X^TY$ have one or more solutions without the assumption that X is invertable?

My only guess is that it has something to do with generalized inverse, but I am totally lost.