It is a matter of perfectly ordinary empirical fact that we live in a universe that, so far as we can tell, is translationally, temporally and rotationally invariant. There is absolutely nothing in Noether's theorem that forbids us from having theories that violate momentum, energy or angular momentum conservation if they happen to describe a universe that has aspects that violate those invariance conditions.
There have been theories that do so, like Dirac's weird large number thing. These are completely legitimate theories, and are entirely subject to absolutely ordinary observational, experimental and inferential verifiability (which is why we know Dirac's large number thing is likely false.)