The storage crunch has reached the memory institutions of the web

The AI data center boom is not only reshaping chip markets and power demand. It is also creating a quieter form of pressure on the institutions that preserve the internet itself. According to archivists and nonprofit operators cited by 404 Media, rising prices and shortages for high-capacity hard drives and solid-state storage are making it harder to maintain and expand digital archives.

The effect is especially acute for organizations that live or die by bulk storage economics. The Internet Archive, Wikimedia, academics, and hobby archivists all depend on a steady supply of drives to ingest, replicate, and protect vast quantities of data. As AI infrastructure scales, they say, those core components are becoming more expensive, harder to source, or both.

Why a storage market spike matters for preservation

Digital preservation sounds intangible, but it is built on physical hardware. Every crawl, upload, backup, and mirrored collection eventually lands on a device that must be purchased, deployed, and replaced. When the price of those devices surges, archives face a direct capacity constraint.

The report offers a stark example: a 2TB external Samsung SSD that cost $159 last fall was priced at $575 at the time of publication. Tracking data from PC Part Picker reportedly shows broad increases across both consumer and enterprise storage beginning around October of the previous year, with some drives doubling in cost or rising by more than 150 percent. In some cases, drives are simply sold out, and a resale market has emerged around scarce models.

The biggest archives are already adapting around shortages

Brewster Kahle, founder of the Internet Archive and the Wayback Machine, described the problem as very real in both time and money. The Archive gathers more than 100 terabytes of new material each day and already maintains more than 210 petabytes of archived material that require ongoing upgrades and maintenance. Kahle said the organization’s preferred 28-30TB drives are either unavailable or priced unusually high.

That kind of demand scale makes substitution difficult. An archive can work around a temporary shortage, but every workaround has tradeoffs in efficiency, acquisition planning, or hardware standardization. For a preservation institution, that can translate into slower growth, more operational complexity, or tighter budget pressure.

Wikimedia sees the same warning signs

The Wikimedia Foundation, which supports Wikipedia and Wikimedia Commons, also told the publication that rising storage prices are a concern. With more than 65 million Wikipedia articles alone and significant media infrastructure behind its broader projects, server and storage capacity are essential operating requirements. Wikimedia said the increases affect not only direct purchases of memory and hard drives but also lead times for server deliveries and the ability to plan future orders.

That matters because nonprofits do not absorb supply shocks the way hyperscale operators can. Large AI companies may help drive demand for storage through data center expansion, but they also typically operate with very different capital flexibility than public-interest archives.

A broader cost of the AI buildout

The story points to a subtle cultural consequence of the AI boom. The same infrastructure race powering new generative systems may also be raising the cost of keeping the web’s historical record intact. That tension is difficult to ignore. While the technology industry pushes to produce ever-larger models and services, the groups preserving the source material of digital culture are being forced to spend more just to stand still.

There is no suggestion here that AI alone explains every storage market movement. But the archive sector is clearly feeling the side effects of a supply environment shaped by aggressive infrastructure demand. When preservation budgets are squeezed, the risk is not abstract. Some data may be delayed, deprioritized, or harder to safeguard at scale.

The internet has always had a memory problem. Links rot, platforms vanish, and formats decay. What makes the current moment notable is that the market for remembering is getting more expensive at the same time the amount worth preserving keeps expanding. For archives, that is not just a procurement headache. It is a structural challenge to the long-term stewardship of digital history.

This article is based on reporting by 404 Media. Read the original article.

Originally published on 404media.co