Not a problem
This is all a self limiting problem. We have the abllity to generate more data now than is possible any use to anyone, this is because it is cheap and easy, but it still has a cost. When the value of that data diminishes below it's cost to create then the problem starts sorting itself out.
The problem at the moment is that we are in transition, there are a lot of people spending a lot of money in a futile attempt to organise, understand and preserve a lot of worthless garbage. Once they grow up enough to realise they are wasting their time (and money) they will eventually stop.
Data of real value will generally be treated well enough. Some will be lost, no big deal, we will always create some more; that data wasn't as valuable as some would like us to believe in the first place.