HDD reliability is only really an issue of running costs for most data centres of any sort. Drives are expected to fail so there's generally several layers of redundancy (Most larger places will generally try to keep data on at least a couple of different geographical locations incase of a major incident at one location).
If a drive sacrifices say 10% of it's running time/reliability for a 20% increase in price/GB and density then it's worthwhile, if it's the other way round(Or equal really given the costs associated with more frequent maintenance) it's not, but that's about as far as the real world effects go.