19:20
@skd geometrically you're looking a map Spec(Z) -> M_{fg} classifying the additive formal group law. as stated, that map isn't flat so the Landweber exact functor theorem doesn't produce a (co)homology theory for you out of it (I don't think integral homology is a functor of MU_*-homology).
one construction of KO comes from a map BC_2 -> M_{fg} which is flat. it doesn't look flat on global sections, but locally it's flat
said another way, KO comes from building KU (which is flat) with a C_2-action. (though Landweber's theorem, and the Hopkins-Miller theorem, aren't actually strong enough to do this)