In fact I recon the ice particles will penetrate quite deeply into simple iron even if they have no significant velocity of their own, but if they do, they will go deep.

The good news is that almost all of the corrosion occurs on external
surfaces that face the direction of motion in some way.  Those are
easily predicted for the design.  Interior components are mostly
shielded since the satellite is moving very much faster than the
average speed of the atoms in the extremely rarefied gas.

I do agree that the actual kilogram amount of corrosion is generally
very, very little though.