Competent engineers in every discipline (HDDs and SSDs alike more than anything) all moved onto the Physics of Failure paradigm ages ago. As a matter of fact, Seagate in the early 2000s understood perfectly well that you could get away with lead-free solder (particularly in BGAs) under the “enterprise” comforts of 24×7 operation with proper cooling, only for it to fail miserably in on/off desktops (and other consumer devices, most-infamously the Xbox 360) – they were an early adopter of lead-free solder on their SCSI drives, but retained leaded solder on their (S)ATA models for the normal time (and thankfully got away from the Barracuda ATA II/III/IV's BGA controller before RoHS, as that combination would have really sucked).
On the same topic, the Barracuda (S)ATA line up to and including 7200.8 was quite euphoric in never willingly cutting any corners, as Seagate could cover any slight losses with their ample SCSI profits (this is the same reason IBM, Quantum and Seagate were able to maintain their quality standards during the industry's 1997/1998 hardships; whereas Maxtor and Western Digital, each with little/no SCSI share and compelled to profit in ATA alone, had no choice but to slack off somewhat). In fact the Barracuda ATA IV (with FDBs as standard) through 7200.8 remain legendary for being extraordinarily reliable in any era – apart from some early defective spindle motors (with manufacturing burrs in the FDB) and substandard STMicroelectronics chips (which Seagate addressed, more‑or‑less, by revising their specifications in late 2004 or early 2005), they almost never fail prematurely without serious provocation; when decently-treated they last well beyond 10 years 24×7 (the POH counter itself being normalized over a 10‑year span, despite the traditional 5 years stated in the manual), and most remain in good (often even perfect) health even today. Even when they've finally had enough of scorching heat and/or excessive humidity (or other corrosive influences), they invariably go out peacefully (with bad sectors and/or degraded read/write performance) – barring operating shock, it's only by grossly exceeding the rated 50,000 start/stops that you'll actually manage to crash their heads…
With Seagate providing an unprecedented enterprise‑standard 5‑year warranty on Barracuda 7200.7s (all capacities) and successors from late 2004 – sadly abandoned already by early 2009 after the vicious Maxtor managers invaded (and overruled Seagate's engineers) – I've followed their example ever since (with the Caviar Blacks and soon-to-be the ST8000DX001) and have yet to be let‑down once. (Make no mistake, the 5400rpm and 3‑year‑warranted regular Red series is consumer‑grade, just like Seagate's U Series 6 was in 2001.) The WD4003FZEX and WD3003FZEX were superb in their own right for finally dethroning Hitachi's “Kurofune” lineage after 9 years, while also idling quieter than 3‑platter Hitachis or Seagates despite being 5‑platter flagships; and now that reliable enterprise SSDs are down to quite reasonable prices (<AU$1/GB for the Exascend SE3 I've used) while even Samsung and Crucial have predictably fallen in the consumer SSD space, I can leave those bad‑old days of the IBM‑reified “consumer”/“enterprise” distinction behind me for good.
The point is, quality hardware is engineered generously enough to reliably last for its rated lifespan (with near‑zero failures), even under worst‑case conditions within the bounds of “normal” use. Consumer stuff has less margin and will foreseeably fail somewhat‑prematurely (though hopefully still‑peacefully) under harsher conditions…
Then there's the “junk” category where even calling it consumer‑grade would be far too nice – completely incapable of lasting a reasonable time, or in the most-cynical cases (let's never forget the DiamondMax Plus 9/MaXLine Plus II and the two following generations; with their blatantly‑falsified start/stop rating, until Apple and other furious OEMs forced Maxtor to fix it in later production) even deliberately designed to catastrophically fail as soon after the warranty‑end as possible.
(Hardware manufacturing ethics really did take a nosedive once IBM's post‑Deathstar FUD caught on – IBM really upped their foot‑shooting game to a .50 BMG – which was probably a factor in ultimately selling their PC business to Lenovo, and despite being a Chinese company, Lenovo's business‑class products largely maintain the honorability IBM was traditionally known for, along with reasonable quality standards.
That same era brought us the CWT‑built Antec PSUs, obliterating their already low‑quality Fuhjyyu capacitors in high heat; and VIA's abysmal KT600+VT8237 chipset, which on top of the VT8237's infamously broken SATA speed negotiation, fell apart with one malfunction after another.
At least I'm glad I only personally suffered through VIA, with my trusty ST380011A unbreakably determined to remain trouble‑free long after the shitbox it came in was a total write‑off…)
The classic Barracudas remain more reliable than any car I know of (even if the Toyota Hilux comes close, and versus Korean models it's no contest); maybe Seagate could break into the EV industry – they'd do an infinitely‑better job of it than Tesla, anyway 😅
