Microsoft's New 10,000-Year Data Storage Medium: Glass (arstechnica.com)
- Reference: 0180828146
- News link: https://hardware.slashdot.org/story/26/02/19/1939246/microsofts-new-10000-year-data-storage-medium-glass
- Source link: https://arstechnica.com/science/2026/02/microsofts-new-10000-year-data-storage-medium-glass/
The system writes data by firing laser pulses lasting just 10^-15 seconds to create tiny features called voxels inside the glass, each capable of storing more than one bit, and reads it back using phase contrast microscopy paired with a convolutional neural network trained to interpret the images. Writing remains the main bottleneck -- four lasers operating simultaneously achieve 66 megabits per second, meaning a full slab would take over 150 hours to write, though the team believes adding more lasers is feasible.
[1] https://www.nature.com/articles/s41586-025-10042-w
[2] https://arstechnica.com/science/2026/02/microsofts-new-10000-year-data-storage-medium-glass/
The cost is the key (Score:3)
If this archives a terabyte cheaply, it will be a wonderful thing! We need good and cheap archival storage and have not had many options for a while now...
Seen this one before. (Score:2)
Reminds me of a prototype laser storage system I saw in 1965 during a guided tour of the Bell labs in Murray Hill. It used lasers to burn spots on a slab of optical material, dont recall any performance or capacity specs -- been a long time. Sure any patents are long expired, but then so is the Murray Hill labs.
Zero obsolescence. (Score:3)
> requiring zero energy to preserve.
Uh huh. Assuming we still have access to a medium reader even 100 years from now. Much less 10,000. Including the native knowledge the read it.
What's it written in? Wait don't tell me. You ironically wrote data on glass in Rust, didn't you? Nerds gonna nerd.
Re: (Score:2)
Presumably in 100 years our AI overlords will be able to take high resolution images of these slabs and instantly decode the contents regardless of the data format while picking through the bones of our collapsed civilization.
Re: (Score:1)
Given that we cannot read material stored in various electronic formats even 50 years ago, this seems like a reasonable doubt. Is there really that much material worth archiving anyway? If you want to last forever, engrave it stone or stainless steel in a human readable format. Then remember that 1,000 years from now, almost no one will care, it anyone is around to read it at all.
*yawn* Vaporware perhaps (Score:1)
I've been seeing stuff about storage tech like this since the Tamarak days in the 1990s with holographic storage. Nothing ever pans out. In fact, we have far less usable storage than the 1990s. Back then, we had floppy, Travan, 4mm, 8mm, DLT, Ditto, HDD, flash (primitive), phase change, and optical.
Now, we have flash drives, HDD and SDD, with LTO being priced out of almost everyone's price range. We have cloud storage, but that's someone else's HDD, SSD, tape or flash drive.
We need to get actual media b
In future Microsoft glass-storage news ... (Score:3)
> though the team believes adding more lasers is feasible
640k lasers should be enough for anyone. :-)
What? (Score:2)
They haven't named it CoPilot Glass Hole Backup yet?
Go figure... (Score:1)
Microsoft would be the ones investigating how well windows can store data.
How dat work? (Score:2)
"each capable of storing more than one bit". I think of a voxel as the smallest usable volume in a 3-D array. How do they get more than one bit? Are they stored in shades of grey? Am I going to have to RTF?
Re: (Score:2)
Probably similar to how multi-level (MLC) NAND flash memory works, but I didn't read the article either