As someone who's made sure to keep the data from my PhD thesis around for almost a decade now, I know that most of it is totally obsolete and will never be looked at by anyone. It is a laudable goal to keep data available, but it'll be a cache with very low hit rate, I think.

However, this was simulated data, it could be recreated by checking out some old version of the code and rerunning it. These days, it wouldn't even take that long. The situation is different where someone's made measurements of the real world, since those are truly irreplaceable.