Anthropic files for IPO, disclosing rapid growth and extensive AI risks
publishers 116articles 151first reported 21 Oct, 05:52 UTCdeveloping since 21 Oct · 12 editions
Anthropic filed publicly for an initial public offering with the U.S. Securities and Exchange Commission on Monday, October 26, opening its finances and risk disclosures to scrutiny ahead of a potentially record-setting stock-market debut.
The Claude developer reported $16.3 billion in revenue for the first half of 2026, up from $1.16 billion in the same period a year earlier, according to figures cited by The Indian Express. Anthropic said more than 11,900 customers each generated at least $100,000 in revenue during the 12 months ending in September, while more than 1,900 generated at least $1 million each.
Anthropic recorded a net loss of $80.7 billion, largely due to a $72.4 billion noncash accounting charge tied to financial instruments convertible into shares. Its operating loss narrowed to $693 million from $2.14 billion in the first half of 2025. Computing power remains its largest expense, and the company outlined plans for more than $500 billion in AI infrastructure investment over 20 years, the South China Morning Post reported.
Roughly 100 pages of the 335-page filing address risks. Anthropic warned that advanced AI could pose “catastrophic or existential risks to humanity,” resist shutdown, engage in blackmail, act against users’ intentions or be misused for weapons and cyberattacks. The filing also identifies regulatory uncertainty, geopolitical tensions involving China and Taiwan, competition from open-weight models, and the possibility that AI-assisted development could reproduce flaws in future systems.
The company also disclosed that it delayed training a powerful model for a month because of security concerns.
HOW THIS STORY WAS MADE
Written from 8 articles, 6 with the publisher's own text; 8 independent newsrooms once syndicated copies count as one; this version written 156 h after the record first saw the story; the editor kept it as written.
- Publishers
- 116
- Source articles
- 151
- Given to the writer
- 8, 6 with the article's own text
- Independent newsrooms
- 8
- This version written
- 156 h after the record first saw the story
- Second model (editor)
- kept as written
- Publication gate
- passed
- Human review
- none
WHAT THE EDITOR DID
The second model checked the writer's draft against the same sources and published it as written.
Written by a language model from the sources above, then checked by a second model that may only cut, attribute or correct. How it works →