Pleroma may save storage by not downloading federated media locally, but hotlinking to 3rd party servers leaves your users vulnerable to tracking by malicious operators, leeches off the bandwidth of remote servers, and, far more practically, does not guarantee that the media that needs to be displayed is in a file format that is playable on the users' devices. For example, as Pleroma makes no effort to convert user-uploaded files to browser-compatible formats, if Mastodon didn't go through the trouble of converting formats, Mastodon users on iOS would not be able to see many "GIFs" and videos uploaded by Pleroma users in the popular WebM format that is not available on iOS (yet).
You may prefer one or the other but there are definitely reasons for why Mastodon does it that way.