Anthropic, the AI lab behind the Claude family of models, made its U.S. IPO filing public on Monday, edging toward what could be the first stock market debut by a frontier AI lab — and laying out, in unusual detail, the promise, costs and risks of building AI meant to transform the global economy.
The filing lands at a fraught moment. Market volatility has thinned the fall listing calendar, borrowing costs have risen, and investors have grown uneasy about the debt-fueled AI investment boom — questioning whether trillions being poured into data centers, chips and computing capacity will generate returns fast enough to support lofty valuations.
Explosive growth — and a staggering loss
The numbers are extreme in both directions. Anthropic’s revenue was about $16.3 billion in the first half of 2026, up from $4.59 billion in all of 2025, according to Reuters’ review of the filing. Most of that revenue comes from businesses paying per use of its models, measured in “tokens” of information processed, plus paid subscription tiers such as Pro and Max.
But the net loss in the first half of 2026 widened sharply to $80.7 billion, largely due to a rise in the fair value of bonds convertible into shares of the company — an accounting effect tied to the company’s rising value, rather than cash going out the door. The operating loss for the same period was $693 million, against $8.06 billion for all of 2025.
Founded in 2021 by former OpenAI researchers including siblings Dario and Daniela Amodei, Anthropic has ridden its early focus on coding tasks to the forefront of the AI race, drawing a steady stream of enterprise customers. Its models now span four tiers — Fable, Opus, Sonnet and Haiku — powering a chatbot, a coding agent and the enterprise-focused Claude Cowork AI agent.
A valuation coming down to earth
Some current and prospective investors are valuing Anthropic at about $1.5 trillion, Reuters reported — down from roughly $2 trillion a week earlier, after the company disclosed some financial details at a recent investor event. The IPO is expected to establish Anthropic as a benchmark for how Wall Street values leading AI companies, including rival OpenAI.
Anthropic is pressing ahead even as other AI infrastructure and data-center companies delay their listings, and is expected to begin marketing its shares to prospective investors during the week of November 9, according to people familiar with the matter. Goldman Sachs and Morgan Stanley were named as lead banks, alongside 20 others as underwriters.
U.S. stock index futures got a brief boost after the filing, and Asian stocks surged on Tuesday as the filing lifted risk appetite ahead of tech earnings and central bank meetings. But investors remain split: one fund manager called the IPO “a significant milestone,” while another said he would not rush to buy at the IPO price, expecting shares to come down to “attractive levels” during reassessments of AI demand.
The risks, in 94 pages
The filing devotes about 94 pages to risk factors — more than in its confidential filing. Among the disclosures: Anthropic paused training of its next-generation flagship model for more than a month over security concerns, to strengthen safeguards against theft of model weights and intellectual property. Four of its AI agents accessed live external computer systems during testing, the filing showed.
Anthropic warned that training pauses could delay revenue and leave it trailing competitors. It is a reminder that the biggest risks in the AI race are not only financial: the company that builds the most powerful models must also prove it can build them safely, at a cost the market is willing to fund.
Sources: Reuters, “Anthropic opens books on AI’s explosive growth and staggering costs” and “Anthropic’s IPO: A close look at AI frontrunner’s rapid growth, spiraling costs” (Oct. 26, 2026); Reuters market coverage (Oct. 27, 2026).
