Maybe in a few years. You mentioned the difficulty of providing enough bandwidth, but you’re talking about the internet and what the narrator described as storage outpacing the capacity growth of bandwidth failing to keep up with demands made to fill that storage.
Mesh networking has its own bandwidth limitations, as well as considerable strengths vs the internet. I could see it assisting with a solution like this particularly in dealing with the last mile problem between user and fiber. But this would assume that traffic was compressed, tailored and filtered to fit the form-factor provided by the very limited bandwidth available on mesh while still keeping it in a physical size and mass given current and near-term technology.
Particularly troublesome here would be the hidden node problem, since one would assume with distributed data solutions that IPFS seems to provide there would be large volumes of requests for pieces of documents would likely produce a significant overhead burden on mesh networks.
Well worth a read and as an aid to better understanding how mesh works, see Ram Ramanathan’s article:
Mesh networking explainers by goTenna’s Chief Scientist!