A dataset at the scale of Scihub would have to pay someone to keep all the various PDFs online (which IIRC is around 70TB by now), which would mean a DMCA or N&T or similar would take down paid IPFS hosters.
I'm not sure how many people would willingly play IPFS hoster for 70TB datasets for free and while not giving a shit about authorities knocking on the door.
I should also add that I am also working on a project to incentivising authors to make their work freely available: https://flockademic.com/
(More info here: https://medium.com/p/the-holy-grail-in-open-access-sharing-t... )