A Seagate IronWolf Pro 16TB lists at $579.99 on Seagate’s own store. The 32TB lists at $1,159.99. Both work out to $36.25 per terabyte.
That is the part worth stopping on. Storage has always been sold on the promise that buying more of it costs less per unit, and at Seagate’s current list prices that promise has flattened to nothing. Twice the capacity, twice the money, no discount for scale. The only reason the 16TB is momentarily the better buy is a 7 percent promotion that brings it to $539.99, which is a sale, not a structure.
There is a tidy explanation available for this, and it is wrong. The tidy explanation is a shortage.
Nothing that goes into a hard drive is scarce
Prices did move, and hard. A tracker run by ComputerBase followed the twelve most popular hard drive models and found they rose an average of 46 percent in the four months from mid-September 2025, with individual models moving between 23 percent and 66 percent depending on the drive (TechPowerUp). The smallest increase in that set was the Seagate IronWolf Pro 16TB, the drive quoted above. The largest was a Toshiba 22TB cloud-scale model.
The same report contains the sentence that makes this story different from every other component squeeze of the last two years. Hard drives, it notes, contain almost no silicon for storage purposes. The platters are made from materials that are not in short supply. Prices rose by close to half “without any substantial supply chain shortage.”
Compare that to the memory and graphics card shortages it is usually filed alongside, where there is a genuine physical bottleneck in wafer capacity. Here there is no bottleneck to clear. Nothing needs to be mined, fabricated or shipped in greater quantity for the situation to resolve. Which means the standard advice that follows every shortage story, wait it out, does not apply, because there is nothing to wait for.
The 2026 supply was sold before 2026 began
What actually happened is an allocation, and the clearest account of it comes from the people doing the allocating. On Western Digital’s second-quarter 2026 earnings call, chief executive Irving Tan told analysts: “As we highlighted, we’re pretty much sold out for calendar 2026. We have firm POs with our top seven customers. And we’ve also established LTAs with two of them for calendar 2027 and one of them for calendar 2028” (Tom’s Hardware).
Seven customers. Not seven thousand. A year’s worth of manufacturing from one of the three companies that make hard drives, committed in advance to a group small enough to fit around a table, with two of them holding contracts that run into 2027 and one into 2028.
The second number from that call explains why nobody at Western Digital is likely to lose sleep over it. The company’s vice president for investor relations, Ambrish Srivastava, said 89 percent of revenue came from its cloud business and 5 percent from consumer. A manufacturer with that revenue mix is not weighing whether to disappoint the small buyer. The small buyer is a rounding error being served by whatever is left over, and has been for a while.
This is the same buildout that is turning up on other lines of the same business’s expenses. We looked at the electricity version of it recently: nearly all of the year-over-year growth in projected peak power demand across the largest US wholesale market is attributable to data centers, which is why a local fight over a single building does not move anyone’s electricity bill. Storage is the same pattern in a different commodity. The demand is enormous, it is contracted years ahead, and it is indifferent to whether you are in the queue.
Moving to the cloud does not get you out from under it
The obvious response is to stop owning drives and rent the capacity instead. It is a reasonable move for other reasons, but it does not dodge this one, because cloud storage providers are precisely the customers who bought the drives.
Backblaze is the useful example here because it has published its own price history. Its B2 storage held at $5 per terabyte per month for seven years, then rose to $6 effective October 3, 2023, a change the company introduced by noting that “after seven years in service with no price increases, the bar was very high for considering any change” (Backblaze). Today the same pay-as-you-go tier lists at $6.95 per terabyte per month (Backblaze). Wasabi, its closest competitor on price, lists at $7.99 (Wasabi).
Seven flat years, then two increases. The assumption that quietly broke is not about hard drives at all. It is the belief, reasonable for most of the last two decades, that storage gets cheaper every year on its own and that a budget line for it can be set once and left alone.
Which businesses this actually hits
Most coverage of a price surge assumes every reader is affected. Here that is not true, and the difference between the two groups is worth establishing before spending anything.
If your data lives on a handful of computers and totals a few terabytes, this barely touches you. Backblaze Computer Backup is $99 per year per computer for unlimited data, with a business tier at the same price that adds multiple users and administrative controls (Backblaze). That is priced per machine rather than per terabyte, so drive costs do not flow through to it in any way you will notice. A business in this position that starts panic-buying drives is solving a problem it does not have.
If you keep tens of terabytes, you are the business this is aimed at. Video production, security camera archives, medical or dental imaging, engineering and design libraries. At $6.95 per terabyte per month, B2 works out to $83.40 per terabyte per year, so 20 terabytes is $1,668 a year, every year. IDrive Team is another option in this range, listing at $11.99 a month for five users and 5TB and $23.99 for ten users and 10TB, with a discount on the first year only (IDrive). Against those figures a $579.99 drive stops looking expensive and starts looking like a capital purchase worth doing the arithmetic on.
Two practical points follow, and both are checkable today rather than aspirational.
First, find out what you are actually storing before you price anything, because most businesses guess high and a few guess catastrophically low. The number is already displayed for you: a Synology NAS shows it in Storage Manager, a QNAP shows it in Storage and Snapshots, Windows shows it under Settings, then System, then Storage, and every cloud provider shows current usage in its own console. Ten minutes of looking replaces a year of assuming.
Second, if you are buying drives, buy the capacity you need rather than the largest available. The per-terabyte advantage that used to justify reaching for the biggest drive is currently gone at Seagate’s list prices, and the largest capacities are exactly the ones the data centers are contracted for. This is the same discipline that applies to buying hardware for AI work, where the honest test is whether you can name the specific job the bigger machine does.
The move that survives either way
Both paths, owning and renting, are now priced off the same constrained input, and both have gone up. The one lever that is not attached to a hyperscaler’s purchase order is how much you keep.
That is less satisfying than a buying recommendation, and it is also the only advice here that does not expire. Retention is the rare cost line a small business controls outright: nobody outside the building decides how many years of raw video, duplicate exports or superseded backups sit on the array. Storage being cheap for twenty years is the reason almost nobody has ever looked. It is not cheap now, which makes this the first good reason in a long time to open the folder and find out what is in it.
One caution before anyone starts deleting. The point of an archive is that it can be restored, and a backup nobody has ever opened is a belief rather than a fact. Trimming what you keep is a good idea; trimming it without confirming that a real file restores from what remains is how a cost-saving exercise becomes an incident.
Frequently asked questions
Why are hard drives so expensive right now?
Because a small number of very large buyers committed to the output in advance, not because of a materials shortage. Western Digital’s chief executive told analysts the company is “pretty much sold out for calendar 2026” with firm purchase orders from its top seven customers and long-term agreements running into 2027 and 2028 (Tom’s Hardware). Prices rose an average of 46 percent in four months from mid-September 2025 with no substantial supply chain shortage behind it, since hard drive platters use materials that are not scarce (TechPowerUp).
Should I buy drives now or wait for prices to come down?
There is no supply shortage scheduled to resolve, and the contracts that absorbed the 2026 supply already extend into 2027 and 2028 (Tom’s Hardware). Waiting is a bet on when a handful of purchase orders end rather than on a factory catching up, so buy on your own replacement schedule rather than on a forecast.
Is cloud storage cheaper than buying hard drives now?
It depends entirely on how much you keep. Backblaze B2 lists at $6.95 per terabyte per month, which is $83.40 per terabyte per year (Backblaze), while Backblaze Computer Backup is $99 per year per computer for unlimited data (Backblaze). For a few terabytes spread across several machines the per-computer price usually wins. For tens of terabytes in one place, buying drives becomes worth costing out.
Are cloud storage prices rising too?
They have already moved. Backblaze held B2 at $5 per terabyte per month for seven years before raising it to $6 in October 2023 (Backblaze), and it now lists at $6.95 (Backblaze), with Wasabi at $7.99 (Wasabi). Cloud providers buy the same drives, so renting capacity changes who absorbs the increase and when, not whether it happens.
