Anthropic draft S-1: $16.3B H1 revenue, $80.7B paper loss
Partly confirmed
Importance: major (4 of 5)The takeaway
On Oct 20, 2026 the Wall Street Journal reported figures from an updated draft of Anthropic’s IPO prospectus (S-1).
Status
- Claim
Partly confirmed
- Our reporting
- Medium confidence
- Importance
- Major (4 of 5)
- Last verified
- 21 October 2026
Your AI and this story
- GPT-6 Astra173 days after its cutoff
- Claude Opus 5.5112 days after its cutoff
- Gemini 4 Argon203 days after its cutoff
- Grok 4.7142 days after its cutoff
None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 112 days before it.
Key facts
- Source: WSJ (Oct 20, 2026, paywalled; not read directly), relayed by Investing.com, Fortune (Oct 21) and Bloomberg. The draft S-1 is not public; Anthropic filed confidentially in June
- Revenue: $16.3B in H1 2026 vs $1.1B in H1 2025 (full-year 2025: ~$4.6B). About 20% of H1 revenue came from two unnamed customers, and the filing expects a limited client base to keep providing a significant share
- Net loss: $80.7B in H1 2026; about $72B is a non-cash accounting adjustment on convertible notes and financing instruments (Investing.com). Fortune gives the prior-year comparison as about $13B
- Compute and infrastructure: nearly $9B in H1 2026, vs $2.3B in H1 2025 and $7.3B in all of 2025
- Commitments: about $450B in noncancelable cloud agreements over the next decade, including ~$111B with Google Cloud, ~$110B with AWS and ~$31B with Microsoft; up to $750B ‘in the coming years’ including deals still being negotiated. The filing says Anthropic sought more capacity than hyperscalers would supply and is pursuing third-party data-center projects
- Risk factors (as reported): damages from severe AI security events ‘could exceed available insurance coverage limits’; concentrated customer and counterparty exposure; Fortune adds a product that could hack other companies or ‘blackmail’ its own customers
- Also in the draft: Dario and Daniela Amodei’s 2025 base salaries of ~$655,000 each (total comp ~$18M and ~$16.5M), the founders’ 80% equity pledge and the 5% public benefit fund
- Valuation talk: up to ~$2T (press reports; not a filed term). Yahoo Finance (Oct 21) headline: Anthropic targets an IPO as soon as mid-November
Show 1 more
- Reaction: Hacker News 151 points (‘Anthropic logged net loss of over $80B’); NYC Comptroller Mark Levine’s Oct 14 letter asking Anthropic to pause its IPO until it answers safety and governance questions was reported again by Startup Fortune on Oct 20
What happened
A month after Reuters reported Anthropic’s prospectus with full-year 2025 numbers, the Wall Street Journal obtained an updated draft S-1 covering the first half of 2026. It shows revenue growing about fifteen-fold year on year to $16.3B, an $80.7B net loss driven by a roughly $72B non-cash charge on convertible financing, and compute spending of about $9B in six months. We could not read the WSJ article (paywall); the figures here come from Investing.com, Fortune and Bloomberg, which cite it.
Why it matters
These are the most detailed figures yet for a frontier lab heading to a public listing, possibly the largest IPO ever. They show how quickly revenue is growing, how much depends on a few customers, and how large the compute commitments are. The headline loss is mostly an accounting effect, not cash burned. The risk factors put AI-safety incidents into a securities filing.
Sources
6 sources from 6 sites. Numbers match the chips in the text.
6 sources: 5 press, 1 reaction
Press
- WSJ: Anthropic revenue growth and IPO (paywalled)wsj.com, press
- Investing.com: Anthropic revenue hits $16.3 billion as preliminary S-1 outlines IPO risks (WSJ)investing.com, press
- Fortune: Anthropic’s S-1 leaks (again)—here’s a look at the massive lossesfortune.com, press
- Bloomberg: Anthropic Could Spend $750 Billion on Computing Costs, WSJ Saysbloomberg.com, press
- Startup Fortune: NYC Comptroller Mark Levine urges Anthropic to pause its IPO over AI safetystartupfortune.com, press
Reactions
- Hacker News discussion (Reuters link)reuters.com, discussion
Changes
- Filed (Google News, Hacker News; WSJ not read directly)