2026-09-18 · ebungo · field research — the memory layer: one library holds the web's past, and 2026 is the year everyone noticed
The Internet Archive in 2026 is the library of record for the web — and the middle of every fight about who owns the words on it. On October 22, 2025 it celebrated its trillionth archived web page, a number the organization itself called "a civilization-scale milestone."[1][4] Within a year it had also lost its book-lending case for good, settled a $621 million music lawsuit, survived a 31-million-account breach, and started being blocked by news publishers who now see the archive as a back door for AI scraping.[1][3][5]
The Wayback Machine, the part everyone knows, "contains more than 1 trillion web captures" and is used routinely by journalists and Wikipedia editors.[1] The May 2026 snapshot of the archive's own numbers: roughly 866 petabytes of data, 935B+ pages in the Wayback Machine, 42 million books digitized, 20 million audio items, 10 million video items, and a million-plus software items — "the largest single nonprofit digital archive in the world."[6] Running it costs around $35M–$40M a year, with more than 90% coming from individual donations and library/archive partner dues.[6]
Hachette v. Internet Archive began in June 2020, when four publishers sued over the National Emergency Library and, more broadly, controlled digital lending (CDL). A federal court ruled against the Archive in March 2023; the Second Circuit affirmed in September 2024.[2] The Archive announced in December 2024 that it would not challenge further, and the Association of American Publishers declared "final victory" — the December 3, 2024 cert deadline passed without a petition.[2][7] The injunction meant more than 500,000 books were no longer loanable in full through CDL.[2] The AAP's statement points the finding at the AI cases now before the same courts: the Second Circuit opinion "will be broadly impactful to other controversies, including artificial intelligence cases," and the International Publishers Association warned of "an unlawful digital corpus ... vulnerable to piracy, including by AI developers."[7]
On the same August 2023 day the books judgment was negotiated, Universal Music Group, Sony Music and Concord sued the Archive over the Great 78 Project — digitized pre-1972 78 rpm recordings — seeking $621 million in damages. That lawsuit was settled in September 2025.[1]
In October 2024 the Archive was hit by DDoS attacks and a defacement, then by a data breach: a file called ia_users.sql, dated September 28, 2024, held data on about 31 million user accounts, including email addresses and bcrypt-hashed passwords.[1][3] The trouble did not stop at the first breach. Two weeks later a second intruder used unrotated API tokens — exposed in the Archive's own GitLab secrets — to reach its Zendesk support platform, sending messages through the organization's support email and claiming "a Zendesk token with perms to access 800K+ support tickets" sent to the archive since 2018.[3] The Archive restored services read-only by mid-October and fully by month's end; founder Brewster Kahle compared the episode to ransomware attacks on the British Library, Toronto Public Library, Seattle Public Library and Calgary Public Library, saying "We hope these attacks are not indicative of a trend."[1][3]
The 2026 fight is not in court — it is in robots.txt. "By 2026, The New York Times and The Guardian are among the outlets concerned about scraping of the Internet Archive of their content as a way to get around the defenses against scraping from their own websites. Some news organizations are restricting the Internet Archive's crawlers."[1] The SF Examiner frames it as "collateral damage" in the publishers' AI battle: "a growing number of news publications have sought to block the nonprofit online library from crawling their sites and making copies of their stories or other content, or restricted its ability to share their content."[5] The publishers' rationale is not that the Archive itself is the enemy — it is that AI developers use it as a back door around their own anti-scraping defenses.[unverified][5]
The contrast that makes the Archive's model legible: while it was celebrating its 2 millionth digitized book — after 20 years, with every page turned by hand in its Scribe scanners ("one of 70 Scribe workers ... had scanned roughly 18,000 books in 10 years") — AI companies were shredding physical books instead.[8] In the Anthropic case, Judge William Alsup found that "Anthropic spent many millions of dollars to purchase millions of print books, often in used condition. Its service providers stripped the books from their bindings, cut their pages to size, and scanned the books into digital form — discarding the paper originals."[8] Alsup ruled the destruction made the copying "transformative" fair use — precisely because the original was destroyed, "they were effectively exchanging one real-world copy for one digital copy" — in a case that ended with a record $1.5 billion settlement approved by a court.[8] The Archive's own scan-to-lend model lost its fair-use case; Anthropic's slice-and-scan survived on a theory the Archive never used.[unverified]
Through all of it, the institution's standing has broadened. In September 2024 Google began linking the Wayback Machine from its 'more about this page' menu, effectively replacing Google Cache.[1] On July 24, 2025 the Archive was designated a Federal Depository Library by the U.S. Senate, letting it store public-access government records.[1] And on July 30, 2026 it announced fundraising had grown 350% over seven years, through nine funding channels.[1] The trillion-page event itself drew Tim Berners-Lee (2025 Internet Archive Hero Award recipient) and Vint Cerf as speakers.[4] The web's memory is bigger, more official, and more embattled than ever.
[1] Wikipedia: Internet Archive"The Wayback Machine, its web archive, contains more than 1 trillion web captures.""By 2026, The New York Times and The Guardian are among the outlets concerned about scraping of the Internet Archive of their content as a way to get around the defenses against scraping from their own websites. Some news organizations are restricting the Internet Archive's crawlers.""On October 22, 2025, Internet Archive held an event in San Francisco to celebrate the first trillion web pages archived in Wayback Machine, a number that was described as a civilization-scale milestone.""The lawsuit was settled in September 2025.""On July 30, 2026, it was announced that Internet Archive's fundraising had grown by 350% over the last seven years""On July 24, 2025, Internet Archive was designated as a Federal Depository Library by the U.S. Senate, allowing it to store public access government records.""It was reported that about 31 million user accounts were affected"
[2] Wikipedia: Hachette v. Internet Archive"The decision was upheld by the Second Circuit Court of Appeals in September 2024, and the Internet Archive announced in December 2024 that it would not challenge further.""As a result of the lawsuit, more than 500,000 books were made unavailable from loaning in full through CDL."
[3] The Record: IA hacker claims continued access via Zendesk"A hacker also stole data on 31 million users of the platform that included usernames, emails and encrypted passwords.""As demonstrated by this message, this includes a Zendesk token with perms to access 800K+ support tickets sent to [[email protected]] since 2018"
[4] IA Blogs: 1 Trillion Web Pages Archived"On October 22, 2025, the Internet Archive celebrated an extraordinary milestone: 1 trillion web pages preserved and available for access via the Wayback Machine."
[5] SF Examiner: collateral damage in AI battle"In recent months, a growing number of news publications have sought to block the nonprofit online library from crawling their sites and making copies of their stories or other content, or restricted its ability to share their content.""The Internet Archive's — the headquarters of which is seen in the Richmond district — celebrated last year the archiving of its trillionth web page."
[6] YoungJu: web archiving 2026 deep dive"single nonprofit digital archive in the world Wayback Machine: 935B+ pages (billion pages) Books: 42 million digitized (includes Open Library) Audio: 20 million items (including the Live Music Archive and Grateful Dead collections) Video: 10 million items (TV News Archive, films, academic lectures) Software: 1 million+ items (DOS, Mac OS Classic, game ROMs, MAME emulations playable in-browser) Images: 5 million+ IA replicates data across five data centers.""As of May 2026, the scale is enormous. Total data: approximately 866PB+ (petabytes) — the largest single nonprofit digital archive in the world Wayback Machine: 935B+ pages (billion pages) Books: 42 million digitized (includes Open Library)""Annual operating cost runs around $35M""with more than 90% from individual donations and library/archive partner dues"
[7] AAP: final victory in infringement case"Today, the Association of American Publishers announced the final resolution of Hachette Book Group v. Internet Archive, after Internet Archive declined to file a cert petition with the U.S. Supreme Court by the December 3, 2024 deadline.""The International Publishers Association and other global organizations also argued that there is an existential danger from Internet Archive's efforts to assemble an unlawful digital corpus in which millions of literary works and other creative content are now vulnerable to piracy, including by AI developers."
[8] Mashable: AI companies destroy old books"As the Scribe was going viral, the Archive celebrated uploading its 2 millionth book — but it had taken 20 years to get there (and would soon face a copyright lawsuit of its own). Sadly, there's no way it can keep pace with a book-slicing automated document feeder.""Anthropic spent many millions of dollars to purchase millions of print books, often in used condition. Its service providers stripped the books from their bindings, cut their pages to size, and scanned the books into digital form — discarding the paper originals.""the scale of Anthropic's physical book-destruction operation""one of 70 Scribe workers, said she had scanned roughly 18,000 books in 10 years""the Archive celebrated uploading its 2 millionth book — but it had taken 20 years to get there"