The digital dark age has already begun. In 2019, MySpace announced it had lost 12 years worth of music and photos, affecting over 14 million artists and 50 million tracks[s]. The data wasn’t stolen. It simply vanished during a server migration, and the company couldn’t be bothered to recover it.
This wasn’t an isolated incident. A 2024 Pew Research Center analysis found that 38% of all webpages that existed in 2013 are no longer accessible today[s]. That’s more than a third of the internet’s historical record, gone in just a decade. The study also found that 54% of Wikipedia pages contain at least one broken reference link, and 23% of news websites have dead links pointing to sources that no longer exist.
The Digital Dark Age Explained
The term “digital dark age” was popularized by information specialist Terry Kuny in 1997. He warned that “we are moving into an era where much of what we know today, much of what is coded and written electronically, will be lost forever”[s]. The irony is brutal: we’re creating more data than any civilization in history, yet future historians may know less about our era than we know about ancient Rome.
The problem isn’t that digital files decay like paper. The problem is that digital files require specific software and hardware to read them. The Dead Sea Scrolls, made of parchment and papyrusA paper-like writing material made from reeds, commonly used for manuscripts in the ancient Mediterranean., remain readable nearly two millennia after their creation. A DVD has an expected shelf life of about 100 years[s]. But even that century assumes you’ll still have a DVD player and software that can interpret the disc’s file format.
When the Cloud Disappears
The shift toward cloud storage has accelerated this problem. When you store photos on Instagram, music on Spotify, or documents on Google Drive, you don’t actually own those files. You rent access to them. And access can be revoked at any time.
Consider what happened to customers of Revolv, a smart home company. One day they woke up to find all their smart home features broken[s]. The company had been acquired by Nest, which decided to shut down Revolv’s servers. Physical devices that customers had purchased became useless overnight because they depended on cloud infrastructure that no longer existed.
A 2022 Harris Poll survey found that 54% of computer owners report having lost data at some point[s]. Of those surveyed, 48% had experienced an external hard drive crash, and 44% lost access to data when a shared or synced drive was deleted. Yet about 20% of people have never backed up their computer at all.
The Warhol Recovery
Sometimes data can be rescued, but it requires extraordinary effort. In 2014, Carnegie Mellon University faculty and students recovered a dozen previously unknown artworks by Andy Warhol[s]. The images had been trapped on aging Amiga floppy disks from 1985. The purely digital images had been sitting in museum archives for nearly 30 years, unreadable because no modern computer could interpret the obsolete file format.
The recovery required forensics experts to reverse-engineer the unfamiliar format, eventually unveiling 28 never-before-seen digital images. Most archives don’t have access to Carnegie Mellon’s computer science department. Most data simply disappears.
What Can You Do?
The digital dark age isn’t inevitable, but avoiding it requires effort. Keep multiple copies of important files across different devices and formats. Don’t rely solely on cloud services. External hard drives fail, but so do cloud providers. The safest approach combines both local and cloud backup.
Consider analog alternatives for truly irreplaceable records. Print important photos. Store critical documents on paper. It sounds old-fashioned, but paper has a proven track record of surviving centuries. Your Google Photos account does not.
The digital dark age has already begun, and the technical mechanisms driving it are more complex than simple data loss. In 2019, MySpace reported losing 12 years of user-uploaded content, an estimated 50 million songs from 14 million artists, during what the company called a server migration[s]. Tech reporter Andy Baio publicly doubted the accidental nature of the loss: “Flagrant incompetence may be bad PR, but it still sounds better than ‘we can’t be bothered with the effort and cost of migrating and hosting 50 million old MP3s.'”
A 2024 Pew Research Center analysis quantified the scope of web decay: 38% of webpages that existed in 2013 are no longer accessible, and even pages from 2021 showed about a 20% inaccessibility rate just two years later[s]. Pooling pages sampled across 2013 to 2023 from Common Crawl, the researchers found that 25% were no longer accessible overall: 16% were individually inaccessible while their root domain remained functional, and 9% vanished because their entire root domain went offline.
Understanding Digital Dark Age Mechanisms
Data preservation failures cluster into three categories: hardware longevity, format accessibility, and comprehensibility[s]. Hardware fails physically over time. Format accessibility breaks when storage media like 5.25-inch floppy disks become physically unreadable by modern hardware. Comprehensibility fails when file formats become uninterpretable by any existing software, even if the raw bits remain intact.
The scale of data accumulation compounds these problems. Global data storage is growing into the hundreds of zettabytesA unit of digital storage equal to one trillion gigabytes, used to describe global or internet-scale data volumes., and Deloitte projects data centers will consume 536 terawatt-hoursA unit of electrical energy equal to one trillion watt-hours, used to measure large-scale electricity consumption. of electricity in 2025, roughly 2% of global consumption, potentially doubling to 1,065 TWh by 2030[s]. This infrastructure requires continuous power, cooling, and maintenance. Any interruption threatens the data it holds.
Cloud Dependency and Service Discontinuation
The subscription model creates structural incentives for planned obsolescenceThe practice of designing and manufacturing products to fail, degrade, or become functionally unusable after a predetermined period, regardless of their physical durability or the manufacturer's technical capability to extend their lifespan.. As PIRG researchers noted, “In order to charge you monthly to activate and deactivate a feature, the appliance has to connect to the internet and run that feature with software”[s]. This architectural choice makes devices dependent on external servers for basic functionality. When Revolv’s servers went offline, purchased hardware became paperweights.
Even institutions dedicated to preservation face legal constraints. In December 2024, the Internet Archive announced it would not pursue Supreme Court review of Hachette v. Internet Archive[s]. The ruling made over 500,000 books unavailable through the Archive’s controlled digital lending program. The organization that preserves over 700 billion webpages through its Wayback Machine now faces restrictions on preserving books.
Format ObsolescenceWhen a file format can no longer be read by any available software, making stored data inaccessible even if the storage medium is intact. and Recovery
The 2014 Andy Warhol recovery illustrates both the possibility and difficulty of rescuing data from obsolete formats[s]. Carnegie Mellon’s Computer Club forensics experts discovered files with names like “campbells.pic” and “marilyn1.pic” on Amiga floppy disks, but the files were stored in a completely unknown format unrecognized by any utility. The team reverse-engineered the format and extracted 28 images, at least 11 featuring Warhol’s signature.
This recovery required institutional resources, expert knowledge, and the physical media remaining intact for 30 years. The Long Now Foundation coined the term “digital dark age” at a 1998 digital continuity conference and has since explored solutions including the Rosetta Disk, a nickel disk with microscopically etched text readable without electronic devices[s]. Microsoft’s Project Silica stores data in quartz glass with potential lifespans of tens of thousands of years. GitHub’s Arctic Code Vault stores repository snapshots in Arctic permafrost.
Mitigation Strategies
The 2022 Backblaze survey revealed concerning backup behaviors: only 10% of users back up daily, while 20% have never backed up at all[s]. Among those who do use backup services, 61% reported low confidence that their backup system actually protects all their data. The survey found that 57% of users who back up to “the cloud” use sync services rather than true backup solutions, a critical distinction since sync services propagate deletions and corruptions rather than preserving historical states.
Technical mitigation requires the 3-2-1 backup strategyA data backup rule: keep 3 copies on 2 different storage types, with 1 copy stored offsite, to protect against data loss from any single failure.: three copies of data, on two different media types, with one copy stored offsite. Beyond redundancy, format migration must occur before media becomes unreadable. The digital dark age isn’t primarily a storage problem; it’s a maintenance problem. Without ongoing effort to migrate formats, refresh media, and maintain access infrastructure, data becomes inaccessible regardless of how carefully it was initially preserved.
