Geocities was the largest experiment in mass amateur web publishing the world has yet seen. At its peak in the late 1990s, the platform hosted millions of personal pages, neighbourhood-organised, with a particular aesthetic that became synonymous with the early consumer web. Yahoo bought the platform in 1999, ran it without significant investment, and shut it down in 2009.

The archival picture in 2026 is genuinely mixed. The Internet Archive captured a substantial fraction of public Geocities pages before shutdown through coordinated crawling efforts. Archive Team, an informal collective of digital preservationists, ran independent crawls that captured additional material. Together these efforts preserved a meaningful but partial record.

What survived: the homepages of users whose pages were accessible through the standard navigation paths. Pages linked from other surviving pages. Pages that received organic search traffic. The aesthetic of the era — the animated GIFs, the under-construction signs, the visitor counters, the MIDI background music — is largely preserved through these archives.

What didn’t survive: pages buried in deep neighbourhood structures that didn’t have inbound links. Pages behind any form of access control. Personal email addresses, guestbook contents, and the many small connections that made Geocities a community rather than just a hosting service.

The recent work on Geocities preservation has shifted from “save what we can” to “interpret what we have.” Several projects are now applying machine learning to the archived pages to reconstruct community structures, identify patterns in user behaviour over the platform’s lifetime, and surface notable individual pages from the larger corpus. The historical research value of this material is genuine and growing.

The cultural importance is perhaps under-recognised. Geocities was the place where millions of people learned that the web was something they could create, not just consume. The HTML and basic CSS skills picked up while building a Geocities page sent thousands of people into early web careers. The social patterns of fan communities, hobby networks, and personal expression online were partially shaped by what the platform encouraged.

The lessons for current platform architectures are uncomfortable. Yahoo’s decision to shut Geocities in 2009 wiped out cultural artefacts that no commercial entity had reason to preserve. The communities that organised around the platform had no recourse. The current generation of platforms (Instagram, TikTok, X, Threads) operate on similar logic. The archives we’ll have of 2020s social media in fifty years will depend almost entirely on whether independent preservationists are crawling now, because the platforms themselves will not.

The Internet Archive’s continued operation through funding pressures and legal challenges is therefore not a niche concern. It’s the closest thing we have to a memory of much of the cultural web. The 2026 Geocities archive exists because someone cared enough to save it. The 2046 Instagram archive will exist if someone cares enough now. The lesson from Geocities is that hoping the platform will preserve itself is naive.

Looking at the surviving Geocities pages in 2026 is an interesting experience. The aesthetic is dated. The content is sometimes embarrassing. The vitality is unmistakable. People believed they were building something for themselves and each other, and that belief shows in the pages in a way that algorithmic feeds rarely capture. Whether that’s a story about the early web specifically or about creative expression generally is a question worth sitting with.