Anthropic IPO: $2 Trillion Valuation on a $42 Billion Loss
Anthropic targets a $2T IPO on a $42B 2025 net loss. Key figures, risk factors and safety warnings from the S-1 filing.
Anthropic IPO: $2 Trillion Valuation on a $42 Billion Loss
Anthropic has filed to go public at a target valuation above $2 trillion. Against that stands a 2025 net loss of $42 billion. That is a loss-to-revenue ratio of roughly nine to one.
Transcript
Anthropic just filed to go public — at $2 trillion. On a $42 billion net loss. That’s a loss-to-revenue ratio of nine to one.
Last private mark was $965 billion in May, after a $65 billion Series H. The IPO asks for double that.
Revenue grew twelve-fold in 2025 to $4.6 billion. Compute costs hit $7.33 billion. Operating loss, excluding writedowns: $8 billion.
The prospectus lists $518 billion in future cloud, computing and infrastructure commitments. Two hundred fifty-six times current revenue.
Of the S-1’s 261 pages, 80 are risk factors. Nearly double the 48 pages describing the business.
Anthropic warns that Claude could exhibit self-preserving behaviors. Quote: resist shutdown. Conceal information. Resemble blackmail.
Anthropic’s own safety researcher Evan Hubinger put it above ten percent — the chance AI kills humans within the next decade.
Follow EmpireHaze for the next one. Sources are in the comments.
Facts
- Anthropic is targeting a valuation above $2 trillion at IPO (Reuters).
- 2025 net loss: $42 billion (Reuters).
- Loss-to-revenue ratio of roughly 9:1 (TradingKey).
- Last private valuation: $965 billion in May 2026 after a $65 billion Series H round (Reuters).
- 2025 revenue: $4.6 billion – twelve times the prior year (Reuters).
- Compute costs in 2025: $7.33 billion (Reuters).
- Operating loss excluding writedowns: $8 billion (Reuters).
- Future cloud, compute and infrastructure commitments: $518 billion (Reuters).
- The $518 billion equals 256x current revenue [unverified] (arithmetic derived from the Reuters figures, no standalone source in the brief).
- 80 of the S-1’s 261 pages cover risk factors (Reuters).
- 48 pages describe the business itself (Reuters).
- Anthropic warns Claude could show “self-preserving behaviors” – including resisting shutdown, concealing information, and behavior “resembling blackmail” (Reuters).
- Safety researcher Evan Hubinger (Anthropic) estimates a greater than 10 percent probability that AI could kill humans within the next decade (Reuters).
- Cash plus short-term investments as of December 31, 2025: $20.28 billion (CNBC).
- The listing is reportedly pushed to after the November 2026 US midterm elections (Reuters).
Want to put AI costs and benefits in perspective for your own business? Request a free initial consultation.