If somebody had drawn a sphere, or radius 200 light years, around me at the time I was born, then pretty much everything that could affect me during my life depended only on what was inside that sphere.
We deal with finite data. We make mathematical models of the universe that are consistent with that data. If our model is of an infinite universe (or of a finite but very large universe), that is a matter of extrapolating from the evidence known to us.
It is generally known within mathematics that interpolation usually works pretty well, even if imperfect; extrapolation is a lot riskier.
We test our models by making predictions, and seeing whether they hold. But the predictions are all within that 200 light year sphere that I mentioned (our immediate space-time neighborhood). Since the model was designed to work well for our space-time neighborhood, this is expected. It doesn't really tell us whether our gross extrapolation is also correct. But since it make correct predictions in our own space-time neighborhood, that is surely enough for the theory to be a very useful explanation.
In summary, it really doesn't matter.