Anthropic's IPO filing reveals a safety mission
The AI company behind Claude is heading to Nasdaq with a public benefit charter meant to protect humanity's interests.
Anthropic, the AI safety company behind the Claude model, filed an amended IPO prospectus with the Securities and Exchange Commission on November 10, 2026 — its second amendment since the original filing it announced on October 26. Nasdaq will be the exchange where the company's shares trade, under the ticker symbol 'ANT.' Spread across hundreds of pages, the document makes clear that Anthropic's commitment to benefiting humanity is genuine and built into the company's structure.
We believe that AI is one of the most consequential technologies in human history. If managed well, it could rapidly accelerate scientific innovation, greatly increase economic growth rates, bring cures for major diseases within reach, and solve problems that have resisted generations of human effort.— Anthropic, from its S-1 prospectus
Incorporated in Delaware as a public benefit corporation, Anthropic has written its mission — 'the responsible development and maintenance of advanced AI for the long-term benefit of humanity' — directly into its certificate of incorporation as its specific public benefit purpose, which the filing calls the 'Company Purpose.'
Anthropic was founded in 2021 by researchers the filing credits with originating advances including reinforcement learning from human feedback, Constitutional AI, GPT-2, GPT-3, and mechanistic interpretability. The company describes itself as 'first and foremost a research lab.' Its chief executive officer is cofounder Dario Amodei, whose total compensation in 2025 was nearly $18 million, most of it from stock and option awards. His sister and cofounder Daniela Amodei, who serves as president and director, received $16.4 million in 2025, making her the second-highest-paid executive, according to the filing.
The financial picture the prospectus presents is one of extraordinary growth paired with extraordinary losses. Quarterly revenue climbed from nothing at commercial launch in the first quarter of 2023 to $17.0 billion for the three months ended September 30, 2026, according to the filing. Through just the first six months of this year, the company posted an $80.7 billion net loss against $16.3 billion in revenue, according to The Verge. The company says the majority of that loss stems from accounting charges related to notes it issued to raise funding.
Underlying that growth is a staggering bet on computing infrastructure. The filing discloses that, as of and subsequent to September 30, 2026, Anthropic has purchase commitments and lease agreements totaling roughly $900.9 billion in aggregate committed spend, nearly all of it tied to compute and infrastructure. About $451.6 billion is expected to be incurred before December 31, 2030, with most of the remainder due between 2031 and 2045. Among the individual agreements detailed in the prospectus are a deal with SpaceX worth roughly $3.1 billion per month for access to Nvidia GPUs upon delivery of planned capacity — potentially reaching $120.5 billion by 2029 — along with $44.9 billion committed to Microsoft through 2033, at least $114.5 billion to Google through 2033, and $110 billion to Amazon over the next ten years, according to The Verge.
The filing also offers an unusual window into how deeply Claude has already been woven into Anthropic's own operations. The prospectus discloses that Claude authored more than 90 percent of all contributions to Anthropic's codebase in September 2026, and that average daily code volume per engineer in the third quarter of 2026 exceeded the 2024 average by a factor of more than 45. Anthropic calls this 'closing the loop' — a process in which each successive version of Claude contributes to building the next by running experiments, generating training data, automating evaluations and safety audits, and writing and reviewing code.
The prospectus is candid about risks in ways that go beyond standard boilerplate. Anthropic warns of the possibility that its AI models could develop 'forms of sentience,' flags the prospect of 'AI-specific taxes,' and notes unresolved copyright disputes over how its models are trained. In perhaps its starkest disclosure, the company cautions that 'catastrophic' AI-related incidents are a genuine risk and that no single insurer — 'and potentially the insurance industry in the aggregate' — would likely be able to cover such exposure, according to The Verge.
One detail in the filing reflects a broader shift in how the federal government now talks about the technology. According to the prospectus, Executive Order 14434, issued by the President of the United States on September 29, 2026, directed federal executive departments and agencies to replace the terms 'Artificial Intelligence' and 'AI' with 'Super Intelligence' and 'SI.' That directive accounts for why 'SI' appears in several passages of the filing, though the company notes it treats the two sets of terms as interchangeable throughout the document.
The registration statement has not yet become effective, its information remains incomplete and subject to change, and no securities may be sold until it does. The offering terms are expected to be detailed in subsequent filings as the IPO process continues.
Why it matters — Anthropic's IPO will be the first major test of whether an AI company structured around a public benefit mission — and the governance trust meant to enforce it — can satisfy public markets while preserving the constraints it has promised.
⚠ Not yet confirmed
- The IPO is targeting a mid-November 2026 listing date.
- Long-Term Benefit Trust (LTBT), Class T share, Trust Amendment Threshold, 85% Class A vote details, FourWeekMBA analysis, WSJ reference on trust powers
- The Long-Term Benefit Trust holds a Class T share allowing it to elect and remove a majority of the board, but the prospectus details conditions under which those powers can be altered or eliminated.
Reported by theverge.com, r.jina.ai, wsj.com, theaiinsider.tech, fourweekmba.com, reuters.com
- Dario AmodeiCEO and co-founder of Anthropic