Redundancy intact (RAID6, hot spare, or a single drive with a backup)? RMA and wait — free is free. Running a degraded single-parity array? Buy now, file the RMA anyway, and keep the returned drive as a cold spare. The live table below shows what buying now costs at each common capacity.
Cheapest in-stock CMR-family drive at each capacity · prices updated every 4-5 hours · the "cost of waiting" column is qualitative by design — see the note below
| Capacity | Cheapest in stock | Cond | $/TB | Buy now costs | RMA costs |
|---|---|---|---|---|---|
| 8TB | Toshiba MG Series 8TB Enterprise SATA Buy → | Refurb | $10.63 | $85.00 today | $0 + weeks degraded |
| 12TB | WD Ultrastar DC HC520 12TB SATA Buy → | Refurb | $19.17 | $229.99 today | $0 + weeks degraded |
| 16TB | Seagate Exos X16 16TB Renewed Buy → | Refurb | $16.87 | $269.99 today | $0 + weeks degraded |
| 20TB | MDD 20TB SAS 12G 7200RPM Enterprise Buy → | New | $15.50 | $309.99 today | $0 + weeks degraded |
The "RMA costs" column is deliberately qualitative. We will not invent a probability that a second drive fails during your wait — an honest figure would depend on your drives' ages, the array's size, and error rates we cannot see. What we state is the mechanism: single-parity arrays run with zero redundancy until the rebuild completes, and the rebuild is the heaviest read the surviving drives will ever face.
An RMA costs nothing and takes time. Buying costs money and takes a day. That is the entire trade, and the variable that decides it is what the time costs you. Community reports commonly put standard RMA turnaround at two to six weeks — attributed as community-reported, since manufacturers publish process rather than guarantees — and Advanced cross-ship RMA compresses the replacement side to days at the price of a card hold. Our RMA guide covers both routes.
For a drive that was not doing anything critical — a spare, a scratch disk, a backup target whose contents exist elsewhere — the answer is easy: file the claim and wait. Nothing is exposed, and free is free. The calculation only gets interesting when something is running degraded, because then the wait is not neutral time; it is time spent one failure away from losing everything the array holds.
The move that resolves it cleanly is doing both. Buy the replacement today at the price in the table above, rebuild immediately, and file the RMA anyway — the recertified drive that eventually arrives becomes a cold spare, which is exactly what you want in a market where replacement stock is not guaranteed to be available when the next member fails. You pay once and get two useful outcomes: an array that stops being fragile now, and insurance against the next event.
Live replacement options · prices updated every 4-5 hours · last checked just now · CMR-family drives only (RAID-safe)
| Drive | Cap | Cond | Warranty | $/TB | Price | |
|---|---|---|---|---|---|---|
| Seagate Exos X16 16TB Renewed | 16TB | REFURB | 5 yr | $16.87 | $269.99 | Buy → |
| WD DC HC550 16TB SATA 7200RPM Enterprise | 16TB | NEW | 5 yr | $20.63 | $330.00 | Buy → |
| Seagate Exos X16 16TB Renewed | 16TB | NEW | 5 yr | $21.87 | $349.99 | Buy → |
| Toshiba MG08 16TB SATA | 16TB | NEW | 5 yr | $23.12 | $369.99 | Buy → |
In a normal market, a spare drive on a shelf is depreciating inventory — you would rather have the cash. The 2026 market inverted that. Our reporting has documented enterprise drive prices climbing roughly 50% into the year and new-production lead times measured in quarters to years; against that backdrop, a drive you already own is worth more than the number on its old receipt, and its real value is that it is here when a member fails rather than four weeks out.
Two live pages sharpen that judgment. Our used-drive value tracker shows what working drives of each capacity actually fetch today — useful both for pricing the spare you are keeping and for deciding whether to sell one you have outgrown. And our price-forecast method frames the buy-now-versus-wait question generally: through this shortage, waiting has consistently been the expensive choice, which is another argument for ending the degraded window with a purchase rather than with patience.
This page argues for buying in the degraded-array case, so it should be equally clear about the cases where it does not. Dual parity, healthy: a RAID6, RAIDZ2 or SHR-2 array that lost one member still has redundancy — the array can survive another failure while you wait, and the urgency drops to ordinary maintenance. Hot spare already engaged: if the array pulled in a spare and rebuilt itself, the failed drive is now just an RMA candidate and nothing is exposed. Backups current and verified: if a total array loss would cost you a restore rather than the data, the calculus changes from catastrophe to inconvenience.
Single drives outside arrays belong in the same category, with one caveat: recover the data before you ship, because the replacement arrives empty and your original is not returned. That rule is universal — it is the first thing our RMA walkthrough insists on, and Seagate states it in its own guidance.
One last framing for the borderline cases. Ask what a total loss of this array would cost — in money, in hours, in things that cannot be re-downloaded — and compare that to the price in the table above. If the drive costs less than an afternoon of your time, the question answers itself; if the array holds nothing you would miss, so does it, in the other direction. The mistake this page exists to prevent is not choosing wrong. It is failing to choose, and letting a degraded array coast for a month because nobody decided anything.
Four-bay NAS, SHR-1, one member dead, backups two weeks old. This is the buy-now case in its purest form. The array has no redundancy left, the backup would lose a fortnight of changes, and the wait is measured in weeks. Buy the replacement at the price in the table above, rebuild tonight, refresh the backup, and file the RMA for the spare. The purchase price is a small fraction of what re-creating two weeks of work would cost, and you end up with a spare you did not have yesterday.
Eight-bay array, RAID6, one member dead, hot spare already rebuilt in. Nothing is exposed: dual parity survived the failure, the spare took over, and the array is whole. File the standard RMA, wait however long it takes, and slot the returned drive in as the new hot spare. Buying here would be spending money to solve a problem that redundancy already solved — the correct answer is patience.
Single external drive holding the only copy of a photo library. Neither option is the first move here. Recovery comes first — the RMA sends your data away permanently, and buying a replacement does not retrieve what is on the failing disk. Get the data onto another drive (or decide on a recovery service if the fault is physical), then handle the warranty at leisure. Our triage hub covers that ordering in detail.
It depends what the wait costs. For a spare drive or a healthy backup target, wait — free is free and nothing is at risk. For a degraded single-parity array, the weeks of zero redundancy are the cost, and a replacement drive is cheap insurance against losing the whole array. The live table below prices that insurance at your capacity.
It becomes a cold spare — arguably the best outcome available. You end the degraded window today with the purchased drive, file the claim anyway, and bank the recertified return for the next failure. In a market where replacement stock is tight and prices have climbed, a spare on the shelf is worth more than it was in cheaper years.
Directionally: much riskier than running healthy, and the risk concentrates in the rebuild. A single-parity array with a failed member has no redundancy left, and the eventual rebuild reads every sector of every survivor — the heaviest sustained load those drives will ever see, at the moment they are least protected. We do not publish a probability figure because an honest one would depend on your drive ages, array size and error rates; we state the mechanism instead.
When redundancy is intact and the data is protected: RAID6 or dual-parity arrays that lost one member, systems with a hot spare that already took over, single drives outside arrays, and any setup with a current, verified backup. In those cases the wait costs nothing but patience, and the free replacement is the correct answer.
Frequently, yes. Arrays accept same-size-or-larger members, and in 2026's pricing the next capacity up often carries better $/TB. Replacing a failed 8TB with a 12TB or 16TB starts a capacity upgrade you were going to want anyway — the upgrade planner prices that path, and the failed drive's RMA return still comes back as a spare at the original size.
Keep it if you have a slot it could fill on short notice; sell it if you are consolidating to larger capacities and it no longer matches anything you run. Healthy drives hold real resale value in this market — our used-drive tracker prices what a working drive of that capacity and age actually fetches today.