Microsoft's new 10k-year data storage medium: glass

Laser writing in glass for dense, fast and efficient archival data storage - https://www.nature.com/articles/s41586-025-10042-w | https://doi.org/10.1038/s41586-025-10042-w - February 18th, 2026

These demos are technically interesting, but often lacking practical rigour.

In 1994 I was hired to design the servo (motor controls) for a holographic storage solution. During the interview I asked a lot of questions, but rarely got more details than "we've got 30+ PHDs, top of their field, trust us it works." They were certain they had it all worked out, and they would be bought or go public relatively soon. I took a chance and joined, which required selling my house and moving from Colorado to Texas. I was the first person to join from the storage industry.

A few weeks in: Me: What are the position holding requirements? PHDs: +/-30um Me: Wow, ummm, how long does that need held? PHDs: milliseconds, maybe hundreds of milliseconds until we get better media. Me: That's impossible

After a few weeks of tolerance analysis on the mechanics of the problem, I proved that the design could never actually work in a product. One off demos on a 6x10ft air table, yes, but integrating that into a full height 5 1/4 form factor could not be done.

They were bankrupt months later. Bell Labs was working on a similar design, similarly confident. They tried to hire me but I declined. They spun out of Bell Labs as In Phase Technologies. Over a period of about 6 years there were press releases, an investment by Steve Jobs, then bankruptcy.

TL;DR Data storage is hard. Getting a consumer level product developed adds much more difficulty than researchers generally understand.Fingerprints, dust, temperature variation, air flow, all create challenges; and removable media creates an even broader set of challenges.