Hmm, all the way up at 1mm. Man, that tape length really adds up, doesn't it? So, to the question exactly as I asked it, I guess obviously yes.

Maybe I can do even better then. If I'm also building the head, the dielectric gap that does the reading and writing can probably be as small as tens of microns, since you can fairly easily make an edge that fine. Keeping the tape clean and even enough to not crash would be hard in that context, though, since it seems like clearance needs to be on a similar magnitude. Does anybody know what consumer tape reader/recorders use?

In tape form getting an even layer of medium might also be challenging, although if I went with a disk instead it could be spin-coated. Highly oriented particles sound doable with a bit of chemistry.

I'd upvote this, but ironically it would make it look like I didn't upvote it.