There's yet not enough devices connected to the internet wiling to share their data. There's a whole lot of companies that need to be created and things to happen (which might never do) for something like a data marketplace to exist. If that ever happens, any company that has the resources to make sense of additional -targeted- streams of data (as they already effectively do) should be able to turn a significant profit from the acquisition of said data. Either for product customization or advertising - or both.

For every transaction a bank processes there's obviously a fixed cost. Once you account for all the costs relative to them in a given year and divide that number with the number of transactions in that same year, I doubt, given the margins of a given transaction, that a bank can break even on 1 cents transactions. That's assuming that the tech the banks use as of now does not improve by magnitudes in a couple years.

> Once you account for all the costs relative to them in a given year and divide that number with the number of transactions in that same year, I doubt, given the margins of a given transaction, that a bank can break even on 1 cents transactions

This is a strange argument. Visa's costs, per transaction, are demonstrably lower than Bitcoin's. Specific banks and systems may be more or less efficient. But fundamentally, the trustlessness that is a blockchain's selling point comes at the cost--by definition in proof of work systems--of efficiency.

intermediaries fees are often more than operational expenses. In some industries, the fees might be significantly high especially when competition is low.

With blockchains it's mostly about operational expenses. Now, bitcoin transaction fee might be currently high but it is a problem of this blockchain implementation and not with blockchain per se.