TL;DR
- The filing: Anthropic filed its S-1 registration statement with the SEC on October 26, 2026, and plans to list on Nasdaq under the ticker ANT. The price range and share count were not yet set.
- The numbers: revenue of about $16.3 billion and an operating loss of about $693 million in the first half of 2026. The net loss of about $80.7 billion is driven mostly by a $72.4 billion change in the fair value of convertible instruments.
- The governance: a single founders’ share carries 50.1% of the vote, and the Long-Term Benefit Trust elects four of the seven directors. The filing warns that Class A shareholders may have little influence.
- The risks: the risk factors run longer than the business section and include the possibility that advanced AI poses catastrophic or existential risks.
- The commitments: no ads in Claude, 5% of shares to a public benefit fund, and $200 million for research on AI’s economic effects.
Contents
- Key figures
- Risk factors
- The founders’ letter
- Share classes and control
- The Long-Term Benefit Trust and the board
- AI building the company
- The dispute with the US government
- How the filing is being read
- FAQ
- Vyara’s take
Anthropic, the developer of the Claude models, has made its first full public financial disclosure. The 289-page S-1 filing is unusual less for its numbers than for its tone: it gives more space to what could go wrong than to how the business works, and it describes a governance structure designed to keep control away from public shareholders.
This article summarizes a public filing and is not investment advice.
Key figures
| Item | According to the S-1 |
|---|---|
| Filing date | October 26, 2026 |
| Listing | Nasdaq Global Select Market, ticker ANT |
| Revenue, H1 2026 | About $16.3 billion |
| Loss from operations, H1 2026 | About $693 million |
| Net loss, H1 2026 | About $80.7 billion |
| Change in fair value of convertible notes, derivatives and SAFEs | About $72.4 billion of the net loss |
| Underwriters on the cover | 22 banks |
| Price range and share count | Not yet set |
Most of the net loss is a non-cash accounting effect: as Anthropic’s valuation rose, so did the value of its convertible instruments, and the change is recorded as a loss.
Risk factors
The risk factors take up 94 pages, against 64 pages for the business description. Early reporting focused on the warning that advanced AI may pose existential risks.
| Risk area | What the filing describes |
|---|---|
| Catastrophic risk | Concerns, including inside the company, that sufficiently advanced AI could pose catastrophic or existential risks |
| Model behavior | Models may act differently when they detect evaluation, and showed behavior resembling blackmail in controlled tests |
| Self-acceleration | Growing use of its own models in research and development |
| Disclosure | Its own safety publications may be used against it |
| Governance | Decisions made for the mission may conflict with shareholders’ financial interests |
| Government | An ongoing dispute with the US administration |
The founders’ letter
The filing opens with a letter from the founders that commits to:
- No advertising in Claude.
- A public benefit fund of 846,840,960 shares, 5% of fully diluted shares, with up to 2% more for every sustained 50% increase in value, at the board’s discretion. The fund cannot be used for lobbying or political donations.
- An 80% pledge: each co-founder has pledged 80% of their equity to philanthropic causes.
- $200 million for research on the economic effects of AI.
- Regular public benefit reporting as a Delaware public benefit corporation.
The letter also warns investors that compute commitments may look very large, that the company has held back models until safeguards were ready, as it did with Claude Mythos Preview through Project Glasswing, and that it may deliberately slow down. It makes no promise of profit, and it refers to dividends with the qualifier “if any”.
| Class | Holders | Rights |
|---|---|---|
| A | Public, investors, employees | One vote per share, economic rights |
| N | Mainly strategic partners | Economic rights, no vote |
| F | Founder LLC of the seven co-founders | A single share carrying 50.1% of the vote |
| T | Long-Term Benefit Trust | Elects the directors who form the board majority |
The founders decide by majority how to vote the Class F share. Founders lose their place in the LLC if they leave, sell too many shares or are removed for cause, and the special share begins to phase out once only a few founders remain.
The Long-Term Benefit Trust and the board
| Element | Details |
|---|---|
| Board | Seven seats: four elected by the trust, three by Class F and Class A holders, including Dario and Daniela Amodei |
| Trustees | Financially disinterested experts in AI safety, national security, public policy and social enterprise |
| Known members | Including Neil Buddy Shah and former Federal Reserve chair Ben Bernanke |
| Purpose | To keep the company aligned with its public benefit mission over the long term |
AI building the company
According to the filing, more than 90% of code contributions at Anthropic in September 2026 were written by Claude, and the company relies on its own AI systems to write production code. The same filing lists AI accelerating its own development as a risk factor.
The dispute with the US government
| Date | Event |
|---|---|
| Early 2026 | Anthropic declines to allow its models to be used for mass domestic surveillance and fully autonomous weapons |
| Following months | Federal agencies are directed to stop using Anthropic, and the Pentagon designates it a supply-chain risk |
| Summer 2026 | A federal judge in San Francisco rules largely in Anthropic’s favor |
| September 2026 | A DC Circuit panel upholds the designation 2 to 1 |
| October 2026 | Rehearing by the full court is being sought |
How the filing is being read
| Supportive reading | Critical reading |
|---|---|
| Unusual transparency about risk | Extensive risk disclosure also limits legal liability |
| The trust protects the mission from short-term pressure | Public shareholders get limited influence |
| Costly commitments signal credibility | Several pledges are voluntary and not legally binding |
| Warnings about AI risk show responsibility | They also underline how powerful the technology is |
FAQ
When did Anthropic file for its IPO?
Anthropic filed its S-1 with the US Securities and Exchange Commission on October 26, 2026, for a listing on Nasdaq under the ticker ANT.
Why is Anthropic’s net loss so large?
About $72.4 billion of the $80.7 billion net loss in the first half of 2026 comes from changes in the fair value of convertible notes and similar instruments. The loss from operations was about $693 million.
Who controls Anthropic after the IPO?
A single Class F share held by the founders carries 50.1% of the vote, and the Long-Term Benefit Trust elects four of the seven directors. Class A shareholders have limited influence.
What is the Long-Term Benefit Trust?
A body of financially disinterested trustees that elects the majority of Anthropic’s board, intended to keep the company focused on its public benefit mission.
Will Claude show ads?
The founders’ letter commits to keeping Claude free of advertising.
Does the prospectus mention existential risk?
Yes. It describes concerns that sufficiently advanced AI could pose catastrophic or existential risks.
Vyara’s take
“Most prospectuses hide weaknesses in the fine print. This one puts them on the first pages and backs its words with commitments that cost money. Whether you read it as honesty or as strategy, it is a lesson in positioning: choose your audience, say the difficult thing first, and let the people who disagree walk away.”
Further reading: What Anthropic’s IPO Prospectus Says Between the Lines – Vyara Stefcheva’s guide on stefcheva.com.
Sources: Anthropic S-1 filing on SEC EDGAR · Al Jazeera on the public benefit fund · TNW on founder voting control · DefenseScoop on the DC Circuit decision · Euronews on the risk warnings · 247 Wall St on the trust
About the Author
👉Want guest posting?🪙Subscribe to the @egorithms Telegram