Laser-engraved ceramic storage device that stores data for 5,000 years targets astounding 100 petabytes per rack by 2030 — 10X performance boost and 100,000 petabytes per rack also on Cerbayte's roadmap
Startup aiming to deliver 100x capacity and 20x speed compared to its pilot system over the next five years.

German storage startup Cerabyte recently presented at the A3 Tech Live conference in Munich, where it detailed an ambitious new roadmap. Blocks & Files has shared a few updated details about how Cerabyte’s plans are expected to unfold over the next half decade as it pioneers innovative ceramic nanolayer-based storage. Highlights included the goal of realizing storage systems that offer 100 petabytes (PB) of storage per rack, with transfer rates of 2 GB/s, by 2030. Capacities could even reach 100,000PB per rack by 2045, should supporting laser technologies advance as expected.
Cerabyte is currently on the verge of completing the deployment of its pilot production system. This is roadmapped for 2025/26, so there are a few months of latitude remaining. However, we are sure that the recently announced partnership with storage stalwart Western Digital will help oil the wheels of progress.
This first system boasts a spacious 1PB storage per rack for the pilot, but only pedestrian 100 MB/s data transfers. Another drawback of these first systems will be their 90-second time to first byte (TTFB) performance. While some of these metrics may seem acceptable for archival storage, given its attractively resilient and durable characteristics (with a touted 5,000-year data integrity), more can clearly be done.
100PB per rack by 2030
Across three generations to the end of the decade, Cerabyte plans to attack some of those specs that are perhaps dulling the shine of its ceramic storage solutions. We may have heard about the firm's planned milestones for CeraMemory and CeraTape before. However, Blocks & Files has some supplementary details that point to other substantial spec boosts coming in parallel.
By 2030, we won’t just see Cerabyte racks scaled up to 100PB, it says. This 100-fold capacity boost will be joined by a 10x performance boost, going from 100 to 2,000 MB/s or more. That’s not all in the performance stakes; the TTFB should also be slashed from 90 to just 10 seconds. Those cartridge-juggling robots are going to get a lot faster…
Cerabyte also discussed crucial factors for enterprises, such as total cost of ownership and the carbon footprint of its systems. On the former, it boasts a significant reduction in TCO over the next five years, dropping from $7,000 to $8,000 to just $6 to $8.
Meanwhile, moving from traditional tape to Cerabyte’s solutions could drop the carbon footprint of global data storage from 2% of global CO2 emissions to nearer 1.25%.
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
100,000PB per rack by 2045
Looking even further ahead on the roadmap has some mind-boggling but ‘far out’ technologies behind the milestones, notes Blocks & Files. For example, by 2045, Cerabyte reckons its femtosecond laser writing tech could be substituted with particle beam matrix tech, reducing write spot size from 300nm to 3nm. This holds the promise of expanding storage on the ceramic media to enable racks with up to 100,000PB.
100,000PB sounds like a lot, but it is estimated that to archive the whole internet in 2025, you would need a spare 181,000,000PB of storage.
Follow Tom's Hardware on Google News to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button.

Mark Tyson is a news editor at Tom's Hardware. He enjoys covering the full breadth of PC tech; from business and semiconductor design to products approaching the edge of reason.
-
rluker5 The storage media will have data integrity for 5000 years and likely none of the 25 readers will work in 25.Reply
How many working bluray players will still be around in 2050 of the millions made to read M-disks?
Once, at an estate sale the ex wife brought me to, I saw and wanted to pick up one of those old wire recorders that used spools of stainless wire, over a mile long, to do the same job as cassette recorders but much earlier. Pre transistor so it was a clunky box of vacuum tubes. It was $10. Ex wife adamantly said "you're not picking up that trash" so I didn't. Sounds like the sort of thing an ex wife would say. But she was right about it's practical useability if wrong about it's entertainment value.
Point being that if someone picked up a mile long spool of wire with writing on it indicating that it was the only remaining copy of some important radio transmission, would it be accessible in 2050?
This exotic, obscure data storage will only be readable until the last reader is lost, after which the data will be effectively unreadable so I say this supposed 5000 year storage is really 25 year storage. -
usertests
Why would a lightly used Blu-ray player break in 25 years? Maybe a failed capacitor? The real problem isn't the longevity of the readers, it's that physical consumer-oriented formats have been consigned to death by Hollywood, the recording and PC industries. So there's no more discs or readers being produced that can replace the broken ones. If we had new readers for a new 120mm 8K or 16K UHD disc, those could read all of the surviving CDs, DVDs, Blu-ray, M-DISCs, etc. Instead, physical media is toast and we have to pirate from streaming services to preserve anything.rluker5 said:The storage media will have data integrity for 5000 years and likely none of the 25 readers will work in 25.
How many working bluray players will still be around in 2050 of the millions made to read M-disks?
Things are very different in the enterprise world. If a new storage solution catches on, there will be companies willing to manufacture units and readers for decades to come as long as someone is willing to pay for it. And there are clearly companies that need to store data indefinitely. Although they will still have to have off-site backups since a 5,000 year ceramic device could still be destroyed by a disaster.
There is a possibility of storage surviving thousands of years into the future while humanity dies, and the readers are also not expected to survive. If you want to provide a solution for this, it would probably involve sealing up some readers so that they can last and at least be examined. You should also etch instructions in stone on how to make a new one, and how to understand various languages. For the benefit of alien visitors. -
rluker5
Try using an old cassette deck. Some parts are plastic and some are lubricated on both those old cassette decks and bluray players.usertests said:Why would a lightly used Blu-ray player break in 25 years?
Edit: Plastic isn't forever. Many common types even mold without sufficient biocide additive.
I used to work in a foundry in the year 2000. That for some reason got a mainframe in the 70s. Glass room thing with the giant reel tapes. Also had a 3.something GB hard drive almost as big as a portable generator with metal disks the size of 33 LPs. There was one guy left who knew how to administrate the thing and keep it running. He also scavenged all of the aluminum cans in the plant for whatever reason. We still used that mainframe for some mold labeling and I had to go over to this ancient terminal that I had to hold down the star key to type a number because it shared the number and letter keys. All I typed on the thing was numbers. But it did have built in solenoids that clicked when you typed a key. I left that foundry and sometime a while back they closed.usertests said:Things are very different in the enterprise world. If a new storage solution catches on, there will be companies willing to manufacture units and readers for decades to come as long as someone is willing to pay for it. And there are clearly companies that need to store data indefinitely. Although they will still have to have off-site backups since a 5,000 year ceramic device could still be destroyed by a disaster.
There's now nobody left to get those mold label records off of that mainframe if they can get it running. When I checked up on that place there was a story about how some vandals broke in. Would they mess with a crappy old mainframe in a glass room in a wood paneled office building? Maybe. Maybe the owner parted the thing off on Ebay prior for a few hundred bucks.
Edit: Turns out that foundry is still open, just looked abandoned in the news story about getting vandalized on the plant floor and in the office. Might even be using that mainframe instead of a 10 minute excel program as well.
100 years for some obscure tech to remain viable in some other form than museum novelty is an eternity. Look around, the storage probably won't be readable in 25 due to lack of interest.
usertests said:There is a possibility of storage surviving thousands of years into the future while humanity dies, and the readers are also not expected to survive. If you want to provide a solution for this, it would probably involve sealing up some readers so that they can last and at least be examined. You should also etch instructions in stone on how to make a new one, and how to understand various languages. For the benefit of alien visitors. -
l3tr1xingame Probably relevant video to the discussionReply
lvEO4IaEJlwView: https://youtu.be/lvEO4IaEJlw -
DougMcC
In 25 years we'll likely be using the final data storage solution (can't get much more dense than single-atom scale storage). So you'll just keep manufacturing readers for that forever. And we'll get better and better at designing them to last, so eventually we should have very durable readers.rluker5 said:The storage media will have data integrity for 5000 years and likely none of the 25 readers will work in 25.
How many working bluray players will still be around in 2050 of the millions made to read M-disks?
Once, at an estate sale the ex wife brought me to, I saw and wanted to pick up one of those old wire recorders that used spools of stainless wire, over a mile long, to do the same job as cassette recorders but much earlier. Pre transistor so it was a clunky box of vacuum tubes. It was $10. Ex wife adamantly said "you're not picking up that trash" so I didn't. Sounds like the sort of thing an ex wife would say. But she was right about it's practical useability if wrong about it's entertainment value.
Point being that if someone picked up a mile long spool of wire with writing on it indicating that it was the only remaining copy of some important radio transmission, would it be accessible in 2050?
This exotic, obscure data storage will only be readable until the last reader is lost, after which the data will be effectively unreadable so I say this supposed 5000 year storage is really 25 year storage. -
thestryker
Seeing as this technology is aimed at the archival/cold storage tape market this wouldn't have to be the case. LTO (which is around 25 years old now) artificially cuts off older revisions due to the degradation of tape which wouldn't be applicable here. In this case they could have newer drives work in a read only manner and keep it going that way.rluker5 said:The storage media will have data integrity for 5000 years and likely none of the 25 readers will work in 25.
If the technology doesn't catch on then it doesn't really matter either way. -
JeffreyP55
Great! I have a good 10 or 20 years left on the old ticker. Sign me up!Admin said:Cerabyte's ambitious plans target 100PB per rack by 2030, rocketing to 100,000PB per rack by 2045.
Laser-engraved ceramic storage device that stores data for 5,000 years targets astounding 100 petabytes per rack by 2030 — 10X performance boost an... : Read more -
Amdlova Not long a go I saved a friend with a floppy drivers... he has some old diskette with some documents on it. Just need 10 driver unit for get one working.Reply
Consumer devices it's developed to be cheap. Not to be durable... -
dimar
In 2050 you'll be able to 3D print a Blu Ray player at home.rluker5 said:How many working bluray players will still be around in 2050 of the millions made to read M-disks?
-
LuxZg This discussion is interesting... And I agree that any 5000 year storage isn't likely to actually be used 5000 years down the line. IMHO, any DIGITAL data should be transferred to a new medium every 5-10 years if you really value it. Size and speed of new storage always trumps old solutions several times over. And while it takes time to move the archived data to new storage medium, it is well.worth it. When floppy went out if general usage, you'd move to CD, after CD you'd move to Bluray, ... Or if you were at LTO you'd move from gen to gen (skipping 2-3 gens as drives would obsolete older gens).Reply