In contrast to backup company Backblaze, here is a report <https://blocksandfiles.com/2023/03/16/failed-disk-drives-fail-after-just-under-3-years-in-use-says-data-recovery-biz/> from one of the ambulances at the bottom of the cliff -- a company that tries to recover data from drives after they have failed. What they find is the average working life of a failed drive is just under three years of total power-on time. (So if a drive doesn’t fail by then, it looks like its chances of surviving longer are increased.) Also worth noting that, when a drive goes bad, it doesn’t go just a little bad (say, one or two bad sectors), it tends to go significantly bad (hundreds or thousands of bad sectors). This has been my experience with bad drives too (at least, those that could still spin up and do some I/O transfers successfully). Interesting quotes: “We found that the five most durable and resilient hard drives from [the two most popular manufacturers, Western Digital and Seagate] were made before 2015. On the other hand, most of the least durable and resilient hard drives from each manufacturer were made after 2015.” and “in general, old drives seem more durable and resilient than new drives,” and “disks with CMR (conventional magnetic recording) appear more durable and resilient than those with SMR.” I think SMR drives are only used for archive purposes, because the process for writing new data onto them is somewhat convoluted and slow. Ironic, though, that an archive drive might not be that reliable ...