Disclosure: BreezyScroll uses Anthropic’s Claude to help draft articles.
Anthropic, the company behind the Claude chatbot, made its IPO filing public on Monday, Oct. 26. Its S-1 registration statement says it wants to list Class A shares on the Nasdaq Global Select Market as “ANT.” The price range and share count are still blank, so there is no IPO price or trading date yet.
The filing also answers a question the big numbers skip: what would a public buyer actually own? Each Class A share carries one vote. But a single share held by the seven co-founders will carry 50.1% of the total vote. A trust will elect four of the seven directors. The S-1 says the setup “may effectively eliminate the ability of holders of our Class A common stock to influence important decisions, including a change in control.”
How an IPO works, and where Anthropic stands
An initial public offering is the first time a company sells shares to the general public. First it files a registration statement with the Securities and Exchange Commission. That document lays out the business, the finances and the risks. Banks known as underwriters then buy the shares and resell them to investors.
Anthropic is at the first public step. A later amendment would fill in the price range, share count and closing date. The company first filed confidentially in June, CNBC reported. That gave the SEC time to review its numbers before Monday’s public version.
Goldman Sachs and Morgan Stanley are the lead underwriters, and 22 banks appear on the cover. Some existing shareholders will also sell stock in the offering. Anthropic will get none of that money.
Its own share of the proceeds is for “general corporate purposes.” The S-1 lists working capital, operating costs, computing power, infrastructure and hiring. Anthropic may also buy or invest in other businesses, but says it has no agreements for any material deal. Some of the money will cover tax bills when employee stock units vest.
Five share classes, explained
Anthropic is a Delaware public benefit corporation. Its directors must weigh shareholders’ financial interests against those of people affected by the company. They must also weigh a stated public purpose: “the responsible development and maintenance of advanced AI for the long-term benefit of humanity.”
That structure shows up in five classes of stock. The table below follows the S-1’s descriptions.
| Class | Who holds it | Votes | Economic rights |
|---|---|---|---|
| A | Public buyers and many existing investors | One per share | Yes |
| N | Mainly strategic business partners | None, except where law requires | Yes; becomes Class A if sold to an unrelated buyer |
| F | One share, held by a Founder LLC of the seven co-founders | 50.1% of the total vote | None |
| T | One share, held by the Long-Term Benefit Trust | Elects four of seven directors | None |
| E | Employees other than the co-founders | Only a few trust-related matters | None |
Employees get one Class E share for every six months worked. They can also earn shares through a donation-matching program.
The trust is meant to have five “financially disinterested” trustees, and its rules bar them from owning Anthropic stock. Its initial directors are Reed Hastings, Vas Narasimhan and Christopher Liddell, with one seat vacant. Class A and Class F vote together on the other three seats. Those will start with Dario Amodei, Daniela Amodei and Yasmin Razavi.
Why public buyers can’t outvote the founders
The Founder LLC’s 50.1% doesn’t shrink when Anthropic issues new shares. So Class A holders “generally will not be able to affect the outcome” of shareholder votes, the S-1 says. That holds “regardless of how they vote.” There is one exception. If most of the founders can’t agree, the Class F share generally votes in line with Class A votes.
The founders’ share has an end point. It starts to unwind if fewer than three founders remain in the LLC for 90 days in a row. A 450-day transition then begins, after which the share is redeemed for nothing. The trust’s share stays in place indefinitely unless set approvals change or redeem it.
Anthropic calls itself a Nasdaq “controlled company.” That status allows exemptions from some board-independence rules, though it plans to use only one, for its nominating committee. The S-1 warns that the multi-class setup may keep the stock out of some indexes. Index funds that track them would then not buy it. It also says Anthropic does “not expect to pay any dividends in the foreseeable future.”
Revenue rose 14-fold. The $80.7 billion loss needs a closer look
Revenue was $16.3 billion in the six months to June 30, 2026. That was up 1,300% from $1.16 billion a year earlier. Full-year revenue was $41 million in 2023, $386 million in 2024 and $4.59 billion in 2025.
Losses grew too: $1.5 billion in 2023, $8.3 billion in 2024 and $42.0 billion in 2025. The first half of 2026 alone brought an $80.7 billion loss.
Most of that is one line. It is a $72.4 billion “change in fair value of convertible notes, derivative liabilities, and SAFEs.” Those are investments that turn into stock later. As Anthropic’s valuation rose, they became worth more, so the company booked them as bigger liabilities. The S-1 says the rise in its convertible notes’ value was the main driver. In plain terms, it reflects the rising valuation rather than money spent running the business, which shows up in the operating loss.
The operating loss, before those items, was $693 million for the half. A year earlier it was $2.1 billion. A separate $8.8 billion charge came mostly from note purchase agreements. Stock-based pay cost another $2.2 billion.
What the risk factors warn about
Anthropic has lost money every year since it was founded and “may not achieve or sustain profitability,” the filing says. It calls the business “highly capital intensive.” One example is a set of deals with SpaceXAI, a SpaceX unit, to spend up to $120.5 billion on computing power through 2029. That should run about $3.1 billion a month once all the capacity arrives, which Anthropic expects in December.
The risk summary also covers competition and reliance on a small number of customers and sales channels. It flags dependence on a few cloud providers that are also investors and rivals. It warns that its models “may behave in unintended, misaligned, or autonomous ways.”
It also describes a fight with the federal government. On Feb. 27 the president ordered agencies to stop using Anthropic’s models. The Department of War also labeled the company a supply-chain risk. Anthropic says it “may experience material revenue losses or business disruptions” as a result. A federal court in California largely ruled in its favor. The D.C. Circuit denied its petition in a 2-1 decision.
The valuation question, and when insiders can sell
With no price set, valuation is the open question. Anthropic was valued at $965 billion in a May funding round. Some investors now value it at about $1.5 trillion, Reuters reported. That is down from roughly $2 trillion a week earlier.
MIT Sloan professor Sinan Aral put that range in context on CNBC. He said $1.5 trillion to $2 trillion is “approximately from the low to high end a 22 times to 30 times revenue multiple.” Axios business editor Dan Primack told CNBC: “They’re looking for scale, they’re going to spend an absolute fortune on compute.”
The listing lands in a wider debate over AI spending, which our explainer on the AI bubble covers. Anthropic also faces copyright suits, including one from Elvis Presley’s song owners.
Insiders won’t all wait the usual 180 days to sell. Some non-founder directors, officers, employees and consultants can sell from the first trading day. More shares unlock 27, 48, 62 and 76 days after the IPO. The next step is an amended S-1 with a price range and share count, which would set up investor meetings and a listing date.