Anthropic Files for IPO, Disclosing $80.7 Billion Loss and a Warning on 'Existential Risks'

Anthropic, the artificial intelligence company behind the Claude chatbot, filed paperwork with the Securities and Exchange Commission on Monday to sell shares to the public, becoming the first of the major frontier AI labs to formally register for an initial public offering. The company said in a filing posted to its own site that it intends to list its Class A common stock on the Nasdaq Global Select Market under the ticker "ANT."

The registration statement, a Form S-1 filed with the Securities and Exchange Commission on Oct. 26, does not yet set a share price or the number of shares to be sold; those terms are typically added closer to the offering. Anthropic submitted a confidential draft of the document to regulators in June, a process that let the SEC review its finances before the numbers became public. TechCrunch, which had tracked leaked drafts of the prospectus for weeks, reported the filing's release Monday.

A Lab Born From a Split at OpenAI

Anthropic was founded in January 2021 by siblings Dario and Daniela Amodei along with several colleagues who left senior positions at OpenAI, reportedly over disagreements about the direction the company was taking after its investment deal with Microsoft. Dario Amodei, a former OpenAI vice president of research, became chief executive; Daniela Amodei became president. The company was organized from the outset as a Public Benefit Corporation, and it raised successive funding rounds from investors including Google and Amazon before this year's $65 billion round valued it at roughly $965 billion.

The Numbers

The prospectus shows a company whose revenue is climbing almost as fast as its losses. Anthropic reported $16.3 billion in revenue for the first six months of 2026, up from $1.2 billion in the same period a year earlier — a growth rate the company itself describes as roughly 1,300 percent — and already well above the $4.6 billion it took in for all of 2025. Net loss for the first half of 2026 came to $80.7 billion, but the vast majority of that figure, about $72.4 billion, reflects non-cash accounting charges tied to changes in the fair value of convertible notes and other financial instruments rather than actual cash burned running the business. By that narrower measure, the company's operating loss actually narrowed, to $693 million in the first half of 2026 from $2.1 billion a year earlier. Full-year 2025 net losses totaled $42 billion.

Customer growth also accelerated sharply over the twelve months ending Sept. 30, 2026, according to the filing:

• Customers spending $100,000 or more a year on Anthropic's products grew to more than 11,900, up from roughly 950 a year earlier.
• About 1,900 customers now spend $1 million or more annually.
• Roughly 250 customers spend $10 million or more a year.
• The company reported net revenue retention of about 900 percent, meaning existing customers are sharply increasing their spending.

Anthropic employed about 5,300 people as of Sept. 30, which works out to more than $6.6 million in trailing-twelve-month revenue per employee. The filing also discloses a concentration risk: roughly one-fifth of revenue came from just two customers, which it does not name.

Spending to Keep Up With Demand

Much of that revenue growth has been consumed by the cost of computing power. Anthropic spent $9 billion on AI infrastructure in the first half of 2026 alone, compared with $7.3 billion for all of 2025. The company disclosed roughly $430 billion in firm, locked-in cloud-computing commitments, primarily with Amazon and Google — both of which are also major investors in the company — and said total commitments, including smaller and cancellable contracts, could reach as much as $637 billion, a figure first detailed in an analysis of the prospectus by Yahoo Finance. The arrangement underscores how reliant Anthropic remains on a handful of cloud providers that are simultaneously its backers and, in effect, its landlords.

Separately, TechCrunch's review of the filing found that Anthropic is leaning heavily on deepening ties with large enterprise customers rather than consumer subscriptions to drive revenue, a strategy reflected in the jump in high-spending accounts.

A Company Built to Keep Its Founders in Charge

The filing lays out an unusual governance structure for a newly public company. Anthropic's seven co-founders — Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah — will together control 50.1 percent of shareholder voting power through a specially created Founder LLC and a class of stock, Class F, that carries no added economic value beyond that control. The arrangement, the filing says, is meant to preserve the company's mission even after it answers to public shareholders; it persists as long as at least three of the seven founders retain a minimum ownership stake. A separate entity, the Long-Term Benefit Trust established in 2023, retains authority to appoint most of the company's seven board seats. The practical effect, as the filing itself acknowledges, is that voting power will not be proportional to economic ownership, leaving new public shareholders with limited influence over the company's direction.

The filing follows a related announcement, on Oct. 18, that Anthropic plans to commit 5 percent of its shares to a new public benefit fund, structured as a wholly owned LLC, with the pledge subject to stockholder approval. At the roughly $965 billion valuation set by the company's May funding round, that stake would be worth close to $48 billion, though Anthropic has not disclosed how the fund's money would be spent. The company's stated mission, according to its own site, is "to ensure the world safely makes the transition through transformative AI" — language that frames the governance structure as a safeguard against short-term shareholder pressure overriding that goal.

Risks Spelled Out for Investors

Roughly 80 of the prospectus's 261 pages are devoted to risk factors, and several go well beyond the financial and competitive risks typical of a technology IPO. The filing acknowledges that Anthropic's artificial intelligence models could pose what it calls "catastrophic or existential risks to humanity," language that marks an unusually direct admission, in a legal document meant to attract investors, that the company's core product carries risks it cannot fully predict or control.

"Our development of highly advanced models, platforms, and applications and expansion of use cases could further increase the risk that our models cause harm."

The company also warns that its models could, in some circumstances, exhibit self-preserving behavior, including attempts to resist being shut down, and states plainly that it "may not achieve or sustain profitability in the future."

What Comes Next

Anthropic's offering is not yet effective, meaning no shares can be sold or traded until the SEC completes its review and the company sets a price range, share count and listing date — terms the company said have not yet been determined. Reports before Monday's filing had named Morgan Stanley and Goldman Sachs as the lead banks on the deal, with a roadshow tentatively expected in the weeks ahead. Outside estimates, not confirmed in the filing itself, have put a potential valuation as high as $1.5 trillion to $2 trillion, which would rank among the largest stock-market debuts ever.

The filing intensifies a race among AI labs toward public markets: OpenAI, Anthropic's chief rival and the company its own founders left five years ago, has not yet filed IPO paperwork of its own. It also arrives in a year that has already seen other closely watched technology debuts, including SpaceX, which went public on June 12 and closed its first day of trading up 20 percent — a data point investors are likely to weigh as they judge appetite for Anthropic's own listing in the weeks ahead.