In mid-August 2026, the Internet Archive suffered a severe power failure at its main facility, which temporarily took down the primary website and the entire Wayback Machine interface. While engineering teams work to fully restore stable access, it highlighted the physical and infrastructure vulnerabilities of hosting a massive portion of the world's digital memory on centralized servers. The whole thing is currently down.
Beyond web archiving, the platform's digital library access remains heavily restricted following major legal losses. Book publishers and major music industry groups successfully sued the Internet Archive over its digital book-lending practices and historical audio preservation projects (like the Great 78 Project). As a result, millions of commercially available books and audio files have been permanently restricted or removed from public web access.
The biggest shift in web access is that major news publishers have begun aggressively blocking the Internet Archive’s web crawlers from taking snapshots of their sites. Hundreds of publications—including The New York Times, The Guardian, The Financial Times, and the USA Today network—have modified their site rules (robots.txt) or implemented hard blocks to prevent the Archive from accessing their journalism.
Publishers fear the Internet Archive is being used as a "backdoor" by tech giants and AI companies. AI agents (like ChatGPT and Claude) have been caught using web archives to scrape copyrighted text for model training and to bypass publisher paywalls in real-time searches.
Because of these blocks, the Wayback Machine can no longer preserve the active historical record of these primary news sites, leading digital rights groups like the Electronic Frontier Foundation (EFF) to warn that we are losing vital pieces of digital history.
News outlets aren't alone; major social platforms have also cut off access. For instance, platforms like Reddit have blocked the Internet Archive from crawling their pages to ensure that any data used for AI training must go through direct, paid licensing agreements with the platform itself rather than a free public archive.