Anthropic is aiming to raise as much as $69.3 billion in its initial public offering by selling 866.7 million new shares at $70 to $80 apiece, according to a copy of a filing seen by Reuters on Monday. At the top of the range the sale would rank among the largest listings on record, and it would make the Claude developer one of the biggest new issues the AI boom has produced.
The arithmetic is straightforward. At $70 the offering brings in roughly $60.7 billion; at $80 it reaches the $69.3 billion headline. That is a step down from earlier expectations of $80 billion to $100 billion, and it lands alongside a separate correction in how investors price the company: valuation targets have slid to a range of $1.1 trillion to $1.5 trillion from estimates that ran as high as $2 trillion in early October, after a selloff in AI-linked equities forced buyers to re-underwrite the sector.
The structure carries a second number that matters more than the headline. Google, Nvidia, AMD, Broadcom, Micron, Microsoft and TSMC will invest a combined $26 billion through a private placement at the IPO price. Those are the companies that supply Anthropic with accelerators, memory, foundry capacity and cloud compute, and the arrangement converts a slice of that supplier base into shareholders at the moment of listing. Anthropic also said cornerstone investors will sell 23.33 million shares in the offering, so not every dollar raised flows to the company.
The filing terms rest on financials that Anthropic disclosed in an amended S-1 on Nov. 2: third-quarter revenue of $16.96 billion, up 47% sequentially, a swing to a $567 million operating loss, and $895.4 billion in computing and infrastructure commitments running through 2045. Read together, the revenue line explains the demand for the deal and the commitments line explains why the company needs the cash.
Timing is now the operative constraint. The 15-day statutory waiting period on the S-1 expired on Nov. 10, which clears the company to start its investor roadshow this week. The Wall Street Journal reported that Anthropic plans to begin trading on the Nasdaq under the ticker ANT on Nov. 19 or Nov. 20, aiming to close before the Thanksgiving holiday on Nov. 26. Pre-IPO perpetual futures tied to the company on Hyperliquid fell about 7% after the terms circulated, a signal that some holders had priced a richer outcome.
The $26 billion placement is the part of the deal worth watching after the first day of trading. Supplier equity in a customer is not new in AI — Nvidia has taken positions across the stack, and Microsoft and Amazon both hold stakes tied to cloud commitments — but a single placement that enrolls seven of Anthropic's most important vendors at the offer price binds the company's capital structure to its supply chain. It also guarantees demand for the deal at a moment when public-market buyers are cautious, which is precisely the kind of support that makes a record-size listing possible and that raises questions about how much of the demand is independent.
What it means: the pricing range, not the model releases, will be the most consequential AI number of the month. Anthropic is the first frontier lab to test public markets at scale, and its clearing price will be read as the market's verdict on whether hundreds of billions in compute commitments can be converted into durable profit. A deal that prices at the top of the range tells the infrastructure chain to keep building. A discount, or a first-day break below the offer price, reprices the entire capital-intensive layer beneath the models.
Comments (0)
Log in to join the discussion
Log InNo comments yet