Yeap. I think Pandas could and should see a lot more use outside data science as performant in-memory data storage in Python.

The best thing is? Most data scientists don't even care about SoA vs AoS - tabular structure is easy to grok, easy to use AND performant by default!