Yeah, there is some popular misunderstanding about what IPFS is... a lot of people seem to think its essentially free or subsidized distributed cloud storage. But the more you dig into it, the more you realize it's just a fairly inefficient caching system.

LOCKSS looks interesting but it seems like it's exclusively for libraries.

I can find no current public document from the Internet Archive explaining what is backed up, where, or at what redundancy level.

From a 2016 blog post:

"Do you do backups too, for example to guard against corrupt data getting mirrored across both copies, or accidental deletion?"

John Gonzalez, the author and IA infrastructure lead, replied:

"We have done experiments to confirm that we can back up large portions of our corpus... but this is not a regular practice for us at this time."

https://blog.archive.org/2016/10/25/20000-hard-drives-on-a-m...