the server to support the hard drives is quite a large part of the total cost, so replacing drives is probably an economical solution to increasing storage
Was going to say the same... A relatively modest server in the past 5 years can handle a few dozen SATA drives without issue. Going from 8tb drives to 16tb doubles capacity. IIRC they tend to cycle out drives around the 5 year mark.
I'm wondering if they've experimented with higher endurance lifecycles risking larger % of failures relying on redundancy and replication to migrate data, or encountered increased OpEx costs.
