On Monday, October 27, Anthropic filed to sell shares to the public for the first time. The company behind the Claude AI system plans to trade on the Nasdaq under the ticker ANT, and reports say the first day of trading could come as soon as November 16.

It has not yet said how many shares it will sell or at what price. Those details are expected in the coming weeks.

The filing is the first time outsiders can see Anthropic’s actual books. In the first six months of 2026, the company brought in $16.3 billion. In the same stretch of 2025 it brought in $1.2 billion. That is growth of roughly 1,300% in a year, which is not a typical pace for a company this size.

The loss figures are where it gets confusing, so it helps to separate them. Anthropic’s operating loss, which covers the real cost of running the business against the money coming in, was $693 million.

That is a big improvement over the $2.1 billion it lost on the same basis a year earlier.

The overall net loss, however, was more than $80 billion. Only a small part of that is cash going out the door. Just under $9 billion came from spending on AI infrastructure and computing power.

More than $72 billion came from an accounting adjustment tied to the value of convertible notes, derivative liabilities and SAFEs, which are the kinds of agreements investors use to put money into a young company in exchange for a future ownership stake.

When a company’s estimated value climbs quickly, those promises become more expensive on paper, and the company has to record that as a loss even though no money has left its bank account. So most of the headline loss reflects Anthropic becoming more valuable, not wasting money.

The company still warns that it “may not achieve or sustain profitability in the future.”

The filing also shows who is paying. Anthropic says it has more than 14,000 enterprise customers. As of the end of September, over 11,900 of them spent at least $100,000 a year, about 1,900 spent at least $1 million, and around 250 spent $10 million or more.

The business depends heavily on companies paying to use Claude, and the filing admits that a lot of revenue comes from a limited number of products and customers. One thing missing is results for the quarter that ended September 30, which the filing does not include.

The valuation talk is big too. Anthropic was last valued at $965 billion after raising $65 billion from investors in May. Reports say the IPO could value it at up to $2 trillion, which would put it among the largest public companies in the world.

Then there is the question of who will actually be in charge. Seven co-founders, Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah, will form a group called the Founder LLC.

That group will hold 50.1% of the voting power, no matter how many shares the public buys.

Separately, a body called the Long-Term Benefit Trust, set up in 2023 to keep the company focused on its public benefit mission, will have the power to appoint most of the board. Anthropic says this setup is meant to protect its AI safety mission.

Critics point out that voting power is not tied to how much of the company the founders own, so ordinary shareholders will have little say in a company that describes its own technology as posing “catastrophic or existential risks to humanity.”

That phrase comes from the risk section, which is where the filing stands out. It runs 93 pages. For comparison, Uber‘s risk section was around 50 pages.

Anthropic sorts the dangers of AI into four groups.

The first is misuse, such as someone using the technology to design bioweapons or carry out cyberattacks. The second is loss of human control. The third is abuse by authoritarian governments, for example to build surveillance states. The fourth is economic disruption.

The company also admits its models might act in “misaligned” ways and take harmful actions that no user or outside party asked for. If one of its models causes extreme harm, Anthropic says it could be held liable for more than its insurance would cover.

On the ordinary business side, the risks are familiar ones: competition from other AI labs, whether it can get enough computing power, uncertain regulation and politics, and the chance that demand for AI grows more slowly than people expect. That last one feeds the fears of an AI bubble.

Rival OpenAI has reportedly pushed its own IPO to next year. So Anthropic is going first, into a market that is still deciding how much it trusts the AI boom.

There is also a tension built into the plan. Anthropic says it believes frontier AI companies must invest in safety and alignment. Once it is public, it will answer to shareholders who want returns, and the company itself acknowledges that safety spending and investor demands could pull in different directions

The founders’ voting control is their answer to that problem. Whether public investors accept that answer is what the IPO will test.