In theory, Pi is used all over the place, and improving its accuracy might give more accurate results; but in practice, we would not be able to verify those results with similarly accurate measurements.

I suspect most of the issue about calculating Pi to large numbers of places is to develop and prove the software concepts (and the maths) that allow you to make these calculations, rather than being too concerned about what the answer actually is.