Key facts
- The record for the biggest stock-market listing now belongs to Anthropic's $81.3 billion initial public offering.
- The offering comprised 980 million shares, up from 825 million initially planned.
- Trading on the Nasdaq under ticker ANT is expected to begin Thursday, November 19.
- Infrastructure partners including Amazon and Alphabet's Google face nearly $900 billion in committed computing spending from Anthropic for model training and operation.
Anthropic sold its shares at $83 apiece on Nov. 18, 2026, in an $81.3 billion initial public offering on the Nasdaq, a record for a stock-market listing. The price came in above the initial range of $70 to $78. Trading on the Nasdaq under ticker ANT is expected to begin Thursday, November 19. The offering size and record status are the headline figures from the debut.
The company plans to balance investor interests with its commitment to develop advanced AI responsibly. The company operates as a public-benefit corporation, like OpenAI, and must consider public-benefit objectives alongside shareholder financial interests. The company has stated it will weigh those interests against one another. The public-benefit model requires balancing financial and societal goals.
The company increased the share count to 980 million from 825 million initially planned. Anthropic itself sold about 956.35 million shares, while existing shareholders offered 23.65 million. The full deal totals $107.3 billion, including a $26 billion private share placement at the IPO price.
Goldman Sachs and Morgan Stanley lead the 22-bank underwriting syndicate, both of which also led SpaceX’s IPO earlier this year. SpaceX’s June offering began at $75 billion and reached $85.7 billion after underwriters exercised their overallotment option.
Revenue reached $33.24 billion in the nine months through Sept. 30, compared with $2.54 billion a year earlier. The IPO filing sets the annualized revenue run rate at $81 billion as of Oct. 31, up $10 billion from Sept. 30. The nine-month revenue figure comes from the prospectus.
Anthropic had 11,900 customers as of Sept. 30, each generating at least $100,000 in annual revenue over the trailing 12 months. The figure stood at more than 1,500 at the end of last year. The customer count measures firms producing at least $100,000 in annual revenue.
The nine-month net loss totaled $104 billion, including large non-operating accounting adjustments. Operating results moved to a $567 million loss from $974 million of income in the prior quarter. The net loss shrank to $23.3 billion from $66.3 billion in the prior quarter because of non-operating items linked to convertible bonds.
The IPO valued Anthropic at about $1.5 trillion on a fully diluted basis, while the non-diluted valuation stood at about $1.1 trillion. Anthropic is the first major developer of frontier AI models to complete a stock-market listing.
Computing and infrastructure spending reached nearly $9 billion in the first half of 2026, roughly four times the prior-year amount. Third-quarter computing outlays of $5 billion reduced the revenue earned for each dollar of computing capacity. The spending covers model training and operation.
Anthropic has committed nearly $900 billion in computing spending for model training and operation with infrastructure partners including Amazon and Alphabet’s Google. About $70 billion of those commitments falls due over the next 12 months. The company has taken on $6.2 billion in debt, which its filings describe as large. The compute commitments are long-term.
According to Anthropic, the computing providers and infrastructure partners rely on credit support and third-party financing to build capacity and acquire chips. The filing warns that contracted capacity might face delays, repricing, reductions or withholding if financing or support becomes unavailable or more expensive.
Anthropic’s IPO filing cautions that building more powerful models and expanding their uses could increase the risk that its models cause harm. Chief Executive Dario Amodei urged the industry to slow development of powerful models, saying AI progress could outpace society’s ability to manage risks. The warning is a safety risk disclosure in the filing, and Amodei’s call adds a public safety stance to the IPO. His comments reflect a safety-focused stance from Anthropic’s leadership.
Anthropic’s Long-Term Benefit Trust can appoint and remove most board members. The trust does not handle daily operations but oversees long-term objectives. These are governance powers over the board, not operational control. Board appointment and removal powers give the trust significant influence over the company’s direction.

The company’s competitors include OpenAI, Google and other AI developers. OpenAI, called Anthropic’s closest rival, may go public by next year. Anthropic’s founders left OpenAI in 2021 to establish the company. Other AI developers also compete in the market.
September data from Ramp cited by Reuters showed OpenAI’s GPT-6 Astra accounted for about 13% of enterprise AI spending, compared with about 8% for Anthropic’s Claude Fable. Capgemini research indicates 59% of executives at large firms plan to use more open-weight models in the coming year because AI costs exceed budgets. The spending figures show how leading providers divide enterprise AI budgets. AI costs exceeding budgets are pushing executives toward open-weight models.
Meta Platforms, described as one of Anthropic’s biggest customers, is reducing its use of the startup’s models while building its own AI capabilities. Demand for Claude models, particularly the Claude Code coding agent for developers, has driven Anthropic’s growth. The coding agent targets developers.
Intensified industry competition has raised questions about Anthropic’s outlook and ability to sustain rapid growth. Armina Rosenberg of Minotaur Capital called the IPO a useful gauge of AI investor appetite and cautioned that one listing will not settle the debate. Rosenberg’s comments offer an investor perspective on the listing, and she frames it as a test of broader appetite rather than a final answer.
Anthropic’s IPO details lifted US stock futures, while Asian tech stocks fell amid wider AI-trade concerns. Chipmakers and cloud companies face pressure from valuation concerns, rising AI infrastructure costs and the sudden exit of Nvidia CEO Jensen Huang. Demand for AI-related stocks remains high.
Other companies waiting to enter public markets may feel reassured if Anthropic’s debut is strong. Regulatory scrutiny of data center construction, electricity use, employment and AI safety could affect Anthropic’s operations. Regulatory scrutiny represents a risk factor for the company’s operations.
Keeping expenses under control will remain important as Anthropic works toward profitability. Baillie Gifford invested roughly $1.5 billion in Anthropic before the IPO.
Dario Amodei, 43, is Anthropic’s co-founder and chief executive, and Daniela Amodei, 39, is its president. Dario worked on AI research and development at OpenAI before leaving in 2021 and studied physics at Princeton and Stanford while also working at Baidu and Google. Daniela worked at OpenAI and Stripe, where she led risk operations, and now manages the executive team across research, engineering, product and commercial work. Both founders worked at OpenAI.
Tom Brown, 39, is Anthropic’s chief compute officer. He led GPT-3 engineering at OpenAI and worked at Google’s DeepMind project, and at Anthropic he oversees the computing resources needed to build and run AI systems. Jack Clark, from Brighton, UK, leads public benefit and the Anthropic Institute; he worked as a technology journalist before OpenAI and later became policy director. Clark’s background spans journalism and AI policy.
Jared Kaplan, 42, is Anthropic’s chief science officer. He is a physicist with a PhD from Harvard known for AI scaling laws research and is also an associate professor at Johns Hopkins University. Chris Olah, 34, is an AI researcher known for interpretability work who worked at Google Brain and OpenAI before joining Anthropic.
Sam McCandlish, 36, is Anthropic’s chief architect. He holds a PhD in physics from Stanford and worked on AI safety research at OpenAI, and he helps shape the technical architecture of Anthropic’s AI systems. The seven founders are former OpenAI employees who left in 2021 to establish Anthropic; they will become billionaires on paper based on the IPO price. Paper valuations of founders’ stakes based on the IPO price range from $19.3 billion to $19.9 billion and do not represent cash received.

Join the discussion