Blockchain as file storage is a fundamentally bad concept because it is incredibly expensive and the ledger provides a poor compromise between redundancy and cost. When you store or backup some data how many copies do you need. The correct answer is one. The only reason we use redundancy in some form is because failure is a possibility. 0 surviving copies is worthless and 1 copy is maximum value. Since we can't know how, when, or why a copy will be destroyed the cheapest form of insurance is to maintain more copies. The problem is that each additional copy provides a smaller incremental gain however the cost of that additional copy increasingly linearly. The blockchain isn't partially redundant it is perfectly redundant for security reasons. If there are 100,000 nodes you don't gain incrementally much more by having 100,000 copies of your data than you do by having say 100 copies but the true cost is 1000x more. That is a bad tradeoff for bulk storage.
The advantage is that due to the incentive structure, if you want to mine Bitcoins you will have to store all data in the blockchain. This means that while any other web service or data storage company may go belly up and shut down servers when they stop getting an inflow of customers, for Bitcoin as long as mining is profitable your data will always be stored on a large number of nodes which have good economic self interest in preserving it. It is really the only service that can offer that sort of economic guarantee.
The number of nodes (100, 100,000, etc) isn't what matters, but rather that Bitcoin nodes are actually getting paid to store data.
A miner of bitcoins, or any derivative thereof that has not been made to do otherwise, can willfully exclude the “null data” type of transaction from the blocks he or she mines and still have the blocks accepted.
. But yes I agree there lies many other benefits in the blockchain that people are not exploiting yet. 