Anthropic, the San Francisco company behind the Claude family of artificial intelligence models, became a public filer on Monday, October 26, 2026, when it posted a Form S-1 registration statement with the U.S. Securities and Exchange Commission. The document had been anticipated for months, and its arrival turns a five year old research outfit into a candidate for one of the largest listings in American market history.
The filing lays out a proposed initial public offering of Class A common stock and an intention to list on the Nasdaq Global Select Market under the symbol ANT. It does not answer the two questions investors care about most: the number of shares to be offered and the price range for the deal have not yet been determined.
Anthropic was founded in 2021 by Dario Amodei and a group of researchers who left OpenAI, the company he had helped build, out of concern that its approach to safety was insufficient. The company now sells Claude through a developer platform and consumer and enterprise apps including Claude Code, Claude Cowork and Claude.ai, and it splits its model lineup into four classes: Mythos, which includes Fable, Opus, Sonnet and Haiku. Customers reach the models directly or through Amazon Web Services, Google Cloud and Microsoft Azure.
The numbers in the prospectus are extraordinary in both directions. Revenue for the first half of 2026 reached $16.3 billion, up from $1.2 billion in the same period a year earlier, a gain of roughly 1,300%. The net loss for the January to June period was $80.7 billion, a figure that dwarfs the scale of the business.
Key Facts
TechCrunch reported on October 26, 2026, that Anthropic's $16.3 billion in first half revenue came alongside an operating loss of $693 million, a sharp improvement from the $2.1 billion operating loss recorded in the first half of 2025. The $80.7 billion net loss was driven mainly by noncash accounting adjustments tied to convertible debt rather than by cash leaving the business, and the company warned in the filing that it may not achieve or sustain profitability in the future.
NPR reported on October 26, 2026, that adjusted operating income reached $1.6 billion in the first half, compared with a loss of $1.8 billion a year earlier. Raj Joshi, a senior vice president in Moody's corporate finance group, described that swing from negative to positive as a strong positive for the AI ecosystem. Anthropic's loss of nearly $81 billion in the six months through June compares with $13 billion in the same stretch of 2025.
CNBC reported on October 26, 2026, that net revenue retention stood at around 900% as of September 30, meaning existing customers were spending about nine times what they spent a year earlier, against a figure of 108% for the median public cloud software company in January, according to Altimeter Capital partner Jamin Ball. The filing also shows more than 11,900 customers spending at least $100,000 on a trailing 12 month basis, 1,900 spending at least $1 million, and about 250 spending at least $10 million. Renaissance Capital reported on October 26, 2026, that those customers include every Fortune 10 company and more than half of the Fortune 500.
Governance is unusual. TechCrunch reported on October 26, 2026, that the seven co-founders, Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah, will hold their shares through a Founder LLC that collectively controls 50.1% of shareholder voting power. The Long-Term Benefit Trust established in 2023 will have authority to appoint most of the board. Because voting power is not proportional to economic ownership, public shareholders would have little practical control over the company.
On the business side, CNBC reported on October 26, 2026, that Anthropic is the largest single customer of SpaceX's AI division and has agreed to spend $3.1 billion per month on compute from facilities in Greater Memphis. Goldman Sachs and Morgan Stanley are lead underwriters. Anthropic booked $19.7 billion in revenue for the 12 months ended June 30, 2026, and filed confidentially with the SEC on June 1, 2026.
Analysis
The headline loss and the headline growth point in opposite directions, and the gap between them is where the real argument sits. A company that lost $80.7 billion in six months would ordinarily be a distressed asset; a company that grew revenue roughly 1,300% while turning adjusted operating income positive would ordinarily be the most coveted growth asset on the market. Anthropic is both at once, and the filing asks public investors to price the second fact while discounting the first.
The bigger picture here is that Anthropic's offering is less a verdict on one company than a referendum on whether public markets will underwrite the capital intensity of frontier AI at the valuations private investors have already accepted. At its last private round in May 2026, Anthropic carried a post money valuation of $965 billion. NPR reported on October 26, 2026, that reports have put the listing at between $1.5 trillion and $2 trillion, which would place it alongside SpaceX, valued at around $1.8 trillion when it listed in June 2026. Renaissance Capital estimated the deal could raise about $75 billion, on par with what it says SpaceX raised in June.
Those figures deserve scrutiny, because the public record is not perfectly consistent. CNBC reported on October 26, 2026, that SpaceX's June debut ultimately raised $85.7 billion, while Renaissance Capital put the SpaceX raise at $75 billion and called Anthropic's estimated deal size roughly equal to it. Either way, the comparison sets an expectation of a record scale debut for a company that is five years old and has never earned an annual profit.
The loss itself is more an accounting story than a cash story, but that does not make it irrelevant. Convertible debt adjustments can swing a net figure by tens of billions without a dollar moving, yet they also signal how much financing structure sits between the operating business and ordinary equity. The more durable numbers are the operating loss of $693 million, the adjusted operating income of $1.6 billion, the 900% net revenue retention and the concentration of revenue among a few hundred very large accounts. A customer base that includes every Fortune 10 company is a genuine moat; a customer base in which roughly 250 accounts spend $10 million or more is also a concentration risk that public shareholders will have to live with.
Why It Matters
For the AI industry, the filing creates the first real public benchmark. Private valuations for model developers have been set by a small circle of late stage investors; a Nasdaq listing produces a daily price that every competitor, lender and customer can see. That matters for OpenAI, which is also planning an IPO but has put those plans on hold, with Sam Altman ruling out a listing this year as ill-advised amid AI safety concerns. It matters for Google and Amazon, both competitors and distribution partners.
For Anthropic itself, going public brings capital for growth and gives existing investors a way to cash out, as Georgetown University markets specialist Jim Angel noted. Angel also warned that there is concern that we may be in an AI bubble. The filing does not hide from that worry: it runs to more than 90 pages of risk factors, lists OpenAI and Google as competition, and discloses billions in commitments with Microsoft.
For regulators and the public, the most consequential disclosure may be the safety language. The S-1 warns that Anthropic's models could develop faster than the company can keep them safe, or could be used to cause harm or take harmful actions on their own, and the governance section acknowledges that the technology poses catastrophic or existential risks to humanity by the company's own admission. Handing public shareholders a non voting economic interest in that risk, while founders retain control, is a structure the market has rarely been asked to price.
Next Up
Renaissance Capital reported on October 26, 2026, that the filing implies a roadshow launch as soon as the week of November 9, with potential pricing the week of November 16. Before then, Anthropic must name a price range and a share count, and it will face underwriters and institutional buyers who will weigh its 900% net revenue retention and $19.7 billion in trailing revenue against a loss nearly five times the size of the revenue base.
The registration statement has been filed but has not yet become effective, and securities may not be sold before that happens. Copies of the preliminary prospectus, when available, will come from Goldman Sachs and Morgan Stanley. Two wildcards remain: whether October's selloff in neoclouds such as CoreWeave and Nebius, which lost more than 30% of their value during the month, deepens before pricing, and whether Federal Trade Commission demands for Anthropic executives to testify in its consumer safety probe turn into a public hearing before the roadshow ends.
Comments (0)
Log in or sign up to leave a comment.
No comments yet. Be the first to share your thoughts.