Cloud got cheap. LTO got faster. The question of whether to archive video to tape in 2026 isn't a nostalgia argument anymore. It's a real cost-and-workflow comparison that swings hard depending on how much footage you're holding and how often you actually pull it back.
Here are the 2026 specs, the 10-year math against cloud, where LTO still loses, and the actual editor profile that should buy a tape drive this year.
LTO-9 in 2026: The Specs That Matter
| Spec | LTO-9 (current) |
|---|---|
| Native capacity per tape | 18 TB |
| Compressed capacity (marketing) | 45 TB |
| Read/write speed | ~400 MB/s |
| Tape cost (mid-2026) | $80-100 each |
| Drive cost (HBA-based, new) | $4,000-5,500 |
| Drive cost (refurb) | $2,500-3,500 |
| Useful life of tape (in storage) | 30 years |
| LTFS (file system support) | Yes, mounts on Mac and Linux |
Native 18 TB is the number to use for planning. The "compressed" 45 TB number assumes textual data with high compression ratios. Video is already compressed (H.264, ProRes, R3D), so the real number you'll see per tape is closer to 16 to 18 TB.
400 MB/s sustained is faster than most external SSDs and significantly faster than HDDs over USB. A full tape writes in roughly 13 hours.
LTFS (the LTO File System) is the feature that took tape from "esoteric IT thing" to "an editor can actually use this." Once a tape is loaded, it appears as a mounted volume on macOS or Linux, and you can drag files in and out of it like any other drive. The illusion only breaks when you realize random access is slow (the tape has to seek).
The 10-Year TCO That Settles the Argument
Here is 50 TB of finished project footage, archived for 10 years, with two retrievals over that period:
| Storage option | Year 1 | 10-year total | Notes |
|---|---|---|---|
| Backblaze B2 ($6.95/TB) | $4,170 | $41,700 | Plus retrieval (mostly free up to 3x storage) |
| AWS Glacier Deep Archive | $720 + retrieval | $8,500 | Includes one full retrieval |
| LTO-9 (DIY, owned drive) | $5,500 | $5,500 to $6,000 | Drive amortized; tape costs negligible after year 1 |
| Cold external HDDs (rotated) | $4,500 | $9,000 | Replace drives every 5 years |
The cliff is real. LTO at scale, over a decade, beats B2 by close to $36,000 on 50 TB. It beats Glacier by ~$3,000 (and Glacier's retrieval ergonomics are worse). It edges out cold HDDs by enough to matter, with significantly better tape longevity than spinning drives sitting on a shelf.
Below 20 TB or so, the numbers favor cloud because the LTO drive itself is too big a fixed cost to amortize. Above 50 TB, LTO is genuinely the lowest TCO option for archive.
Where LTO Still Loses
The TCO math is only part of the story. Tape has real downsides that you have to budget for.
Retrieval friction. Pulling a single file from a tape involves: finding the right tape, loading it into the drive, mounting LTFS, waiting for the seek, copying the file off, ejecting, and re-shelving. The file you want shows up in 5 to 30 minutes depending on where it sits on the tape. With B2, the same file is in your downloads folder in seconds.
The drive is a single point of failure. LTO drives have heads that wear out. A new drive is rated for tens of thousands of hours of operation, but you only have one. If it dies before you've migrated to LTO-10, you're hunting for a replacement that can read your tapes. (LTO-9 drives can read LTO-9 and LTO-8. LTO-10 drives will read LTO-10 and LTO-9.)
No off-site without physical shipping. Cloud archives are inherently off-site. Tape archives are off-site only if you physically move tapes to another location. This is solvable (a fireproof safe at a partner's office, or a cloud-of-tapes service like Iron Mountain) but it's a workflow you have to build and maintain.
Verification takes time. Tape can develop bit errors over years. Best practice is to do a verification read of each tape every 6 to 12 months, which means dedicated read time on the drive. For a 50 TB archive across 3 tapes, that's about 40 hours of drive time per year just for verification.
No casual sharing. You can't send a client a link to a file that's on tape. Anything you might share, you have to keep on a faster tier.
The Profile That Should Actually Buy LTO in 2026
Don't buy LTO if you have less than 30 TB of archive. The drive cost won't amortize. Stick with B2 or cold HDDs.
Don't buy LTO if you pull files from your archive monthly or more. The retrieval friction will drive you crazy and you'll end up keeping a parallel copy on B2 anyway.
Don't buy LTO if you're a freelancer working out of coffee shops. The drive is a chunky thing that lives in your office. If you don't have an office, this is the wrong tool.
Buy LTO if:
- You have 50+ TB of finished project archive
- You retrieve files less than a few times a year
- You have a permanent workspace where the drive lives
- You're committed to running a tape lifecycle (verification, eventual migration to LTO-10, off-site rotation)
- The 10-year cost savings are large enough to justify the workflow overhead
For the right shop, LTO is the cheapest archive that exists, by a wide margin. For the wrong shop, it's an expensive paperweight.
The Hybrid Most Editors Actually Want
Pure LTO is the right answer for archive-only workflows. Most editors want a hybrid: LTO for the deep cold archive, B2 or Wasabi for the most-likely-needed slice.
A typical hybrid for a 50 TB shop:
- LTO-9 master: full archive, 3 tapes, kept in a fireproof safe ($5,500 in year one)
- Hot copy on B2: most recent 5 TB, immediately accessible (~$35/month)
- Total 5-year cost: roughly $7,600
Compare this to pure B2 at $20,850 over five years. The hybrid is closer to a third of the cost while preserving fast access to the files you'll most likely actually need.
The calculator can model your specific volume and retrieval patterns to find the breakeven point between hybrid and pure-cloud.
Strip Before You Tape
Here is the part that nobody tells you about LTO: every byte you write costs both write time and tape capacity. A typical 50 TB "wrapped" archive on a freelance editor's drive contains 30 to 60 percent media that was imported into a project and never made it onto a single timeline. Cache files, proxies, render previews, and source clips that were considered and dropped.
If you're about to write 50 TB to tape over 13 hours per tape, finding out 20 TB of it is dead weight saves you a tape, an hour of write time, and the storage in your fireproof safe. The grip truck doesn't load gear that's not on the lighting plan. Your archive shouldn't either.
Clip Sweeper was built to find this. It scans Premiere projects, walks the actual edit reference chain, and tells you what's used in any timeline versus what's just sitting in a bin. Running it before an LTO archive is the single highest-value cleanup step you can do, because every minute of LTO write time and every dollar of tape is a real cost.
LTO is genuinely worth it in 2026, for the right scale and the right workflow. Run the math, audit before you write, and you'll have an archive that pays for itself ten times over.