|
Getting your Trinity Audio player ready...
|
For the past several years, my workflow has been built around a stack of SanDisk, Crucial & Samsung T7 portable SSDs — 1TB, 2TB, and 4TB. I bought them because they were fast, small enough to live in a pouch in my bag, and reliable enough to trust with client work. Here’s the pipeline I’ve been running:
- Ingest RAW NEF files from my CFexpress 4.0 Type B card into Photo Mechanic Plus, writing directly onto the external SSD.
- Process the RAWs and export JPEGs to that same SSD.
- Upload the finished JPEGs to PhotoShelter, where I have an unlimited storage plan, for client delivery.
- Copy both the RAW folder and the JPEG folder from the SSD over to my NAS for long-term storage.
That last step is what got me thinking. I’d been treating those SSDs like they were “getting worn out” from all that ingest-process-delete cycling — like I was using up some finite resource every time I filled the drive and cleared it again. Between that assumption and the sticker shock of replacing them, I started looking at a different long-term storage solution, and what I found changed how I think about the whole setup.
Why SSD prices went haywire
If you’ve priced out an SSD recently and did a double take, you’re not imagining things. A wave of enterprise and AI data center buying has been consuming NAND flash production capacity throughout 2026, and it’s dragged consumer SSD pricing along. TrendForce reported enterprise SSD contract prices rose roughly 80% in the first quarter of 2026 alone, as AI infrastructure demand outpaced NAND output, with enterprise SSD revenue jumping over 86% quarter-on-quarter while supplier inventories fell to historic lows.
It’s not just an enterprise problem trickling down slowly — it’s reshaping the whole market. Cloud providers are locking in long-term supply of server-grade DRAM, high-bandwidth memory, and enterprise SSDs, and with a large share of wafer capacity committed to those uses, less is available for consumer drives, pushing retail prices up across the board. One industry analysis noted that in the consumer market, a standard 1TB SSD that was affordable in late 2025 had nearly doubled in price by early 2026.
The forecast isn’t encouraging for anyone hoping to wait it out. Analysts generally agree that meaningful price relief isn’t likely before 2027–2028, since new NAND factory capacity won’t come online in volume until then, even as AI-driven demand is expected to moderate somewhat after 2027. One tracker described contractual NAND prices as having multiplied four to four-and-a-half times over nine months, with no meaningful drop expected before late 2027 at the earliest.
I watched it happen in real time with my own gear. I paid around $100–150 for my Samsung T7 SSDs when I bought them. Today, that same 2TB T7 runs $450–500 at major retailers — more than triple what I paid.
What sent me looking for an alternative in the first place
I never thought my T7S were worn out. Quite the opposite — they were inexpensive when I bought them, and that made them look like a perfectly reasonable place to just let my image library live long-term instead of buying something else for that job. If a drive is fast, reliable, and cheap enough, why not let it double as long-term storage?
It was only the price jump — watching a 2TB T7 go from $100–150 to $450–500 — that sent me looking at alternatives at all. But once I started researching whether SSDs actually make sense as long-term storage, I learned something that had nothing to do with price and everything to do with how flash memory behaves once you stop actively using it.
SSDs aren’t really built to sit idle. NAND flash stores data as an electrical charge trapped in each cell, and that charge is meant to hold steady during regular use — the controller runs background checks, corrects small errors, and refreshes weak cells every time the drive powers on. Pull the power and park the drive in a drawer, though, and none of that maintenance happens. The charge slowly leaks out of the cells over time, and depending on the NAND type, that can mean real data loss.
That retention window is shorter than most people assume, and it varies by NAND type. Industry data-retention guidance generally puts QLC NAND (used in many budget drives) at roughly 1–3 years unpowered before cells are at meaningful risk, TLC NAND (what most consumer drives, including the T7, use) at around 3 years, and older MLC and SLC NAND at 5 and up to 10 years respectively — with retention shrinking further the more the drive has already been written to and the warmer it’s stored. One long-running independent cold-storage test of consumer TLC SSDs found that drives which had already been heavily used showed signs of corruption after just two years sitting unpowered.
Spinning hard drives don’t have this particular failure mode. They store data magnetically rather than as an electrical charge, so a mechanically sound HDD doesn’t lose data just from sitting on a shelf — multiple decades of unpowered retention is a reasonable expectation, which is a big part of why archivists and NAS-based backup strategies have leaned on spinning disk for so long.
So the T7s were never “worn out” — but they also weren’t really built for the job I was quietly asking them to do. A drive I fill for a shoot or two and then let sit for months as my de facto long-term copy is exactly the scenario where SSD data retention is weakest. That realization, plus the price jump, pushed me to add a spinning drive to the mix instead of just buying another SSD.
So why add a spinning hard drive at all?
Between the retention issue and the price, an HDD made sense for the long-term-storage layer specifically — and there’s a mismatch worth calling out between what I actually need from that layer and what I was paying for.
I bought an Avolusion PRO-G1 Series 14TB (16TB drive, ~14TB usable after formatting) USB 3.2 Gen 2 (10Gbps) external hard drive with a built-in USB hub, for $360. Today, a single Samsung T7 2TB SSD costs roughly $450 — meaning one 2TB SSD now costs more than a 14TB spinning drive. That’s about $225/TB for the SSD versus roughly $26/TB for the HDD, a gap that simply didn’t exist when I bought my T7s a few years ago.
The math works in the HDD’s favor right now because the NAND shortage is squeezing SSD and HDD pricing very differently. One industry tracker noted that hard drive prices climbed too during this shortage, but from a far lower base — so the price-per-terabyte gap between enterprise flash and spinning disk actually widened rather than closed. There’s also a supply-side twist: some of the same AI infrastructure buildout squeezing SSDs is also straining nearline HDD supply, as a severe HDD supply shortage has pushed data centers to substitute high-capacity SSDs for HDDs, which is part of what’s driving SSD demand even higher. Consumer-grade external HDDs like the Avolusion haven’t been swept up in that enterprise scramble the same way, which is why pricing still looks normal.
And when I looked honestly at how I actually use that 14TB drive, a spinning disk made even more sense. It isn’t the drive that leaves the house with me, and it isn’t my only copy of anything — it’s one leg of a three-legged stool alongside my NAS and PhotoShelter’s unlimited-storage plan. I don’t need SSD speed or SSD shock resistance for a drive that sits on a desk and holds a third copy of already-safe files. A 7200 RPM HDD is plenty fast enough for that job, and it’s dramatically cheaper per terabyte right now.
My adjusted workflow
I’m not abandoning SSDs — for active ingest and processing, where read/write speed actually matters while I’m working, the T7s stay in the loop exactly as before. What’s changing is the fourth leg:
- Ingest RAW NEF files from the CFexpress card into Photo Mechanic Plus, onto an SSD.
- Process and export JPEGs to the same SSD.
- Upload JPEGs to PhotoShelter for client delivery.
- Copy the RAW and JPEG folders to the NAS — and now also to the Avolusion 14TB drive as an added layer, before deleting anything off the SSD.
- Clear the SSD for the next job, knowing the images now exist in three places, one of which cost a fraction of what SSD storage costs today.
Whether this stays permanent or reverts once SSD pricing normalizes, I honestly don’t know yet—that depends on how the NAND market shakes out over the next year or two, and how comfortable I stay with spinning disk as a working tier rather than just a backup tier. But for now, it’s an extra safety net that costs less than a single 2TB SSD, and it bought me time to let my existing T7S keep earning its keep instead of rushing to replace gear that, it turns out, wasn’t nearly as worn as I thought.
The takeaway for fellow photographers: before you assume your SSD is on its last legs because you’ve filled and cleared it a hundred times, look up what TBW actually means for your drive and do the math on your real write volume. You may have a lot more runway than you think — which matters a great deal in a market where that same drive now costs three times what you paid for it.
