The biggest problem with this idea is proving data replication and guaranteed retrievability to the users in a trustless manner. With a traditional cloud provi

The biggest problem with this idea is proving data replication and guaranteed retrievability to the users in a trustless manner. With a traditional cloud provider, you trust them, upload your data and they deal with the rest, backing it up internally and guaranteeing retrievability with their reputation. But with blockchain p2p storage, it's really hard to provide the same guarantees. If you only upload the file once and let the blockchain replicate it, then a malicious (or greedy) entity can manipulate the blockchain and control all replicas (backed by a single disk copy). So your software must encrypt multiple copies of your data and upload seperate replicas with different encryption keys. But that still leaves you open to a very malicious entity going out of it's way to track all uploads from a node, and monopolising control over them, then deleting them when you try to download them. To actually provide the same guarantees that a cloud provider does in a trustless way, your local software will also have to continually download chunks to make sure they are still retrievable, and you might also need to use the tor network for all accesses. With all that extra complexity, there is a lot to go wrong, and a traditional cloud provider might end up being the cheapest option. I used to agree with you, but after trying to design such a system with these trustless guarantees myself, I've come to the conclusion that there is no good use case for blockchain, outside of exchanging value.