Anthropic, the company behind the Claude AI models, has filed to list its shares on the US stock market under the ticker ANT, after weeks of leaked versions of its filing. It reported revenue of $ 16.3 billion for the first half of 2026, about R268 billion, up from $ 1.2 billion (roughly R20 billion) in the same period last year. It has not yet said how much it plans to raise or at what share price.

Growth was not the same as profit. Anthropic recorded an operating loss of $ 693 million, about R11.4 billion, in the first half, smaller than the $2.1 billion loss a year earlier. Its net loss was far larger, at $80.7 billion (~R1.3 trillion), but the filing says that was mostly driven by accounting adjustments tied to convertible debt rather than cash spent. The company states plainly that it "may not achieve or sustain profitability in the future".

Its business is heavily built on large companies. By the end of September, more than 11,900 customers were spending at least $100,000 a year, about 1,900 were spending $ 1 million or more, and about 250 were spending $ 10 million or more. More than half of those $100,000-plus customers now use more than one of its products, such as the Claude platform alongside apps like Claude Code, up from about 30% at the end of last year, and existing customers were spending about nine times as much as a year earlier.

The filing also names two risks. Amazon, Google and Microsoft, which sell Claude through their cloud services, increasingly see Anthropic as a competitor and may prioritise their own needs for computing power, and losing or shrinking one of its largest customers could hurt its business. The structure after listing is the most debated part. Anthropic's seven co-founders, including chief executive Dario Amodei and president Daniela Amodei, will form a group called Founder LLC that controls 50.1% of shareholder voting power.

Its Long-Term Benefit Trust, set up in 2023 to protect its public benefit mission, will appoint most of the board. Anthropic says this is meant to preserve its focus on AI safety. The listing follows its pledge last week to put 5% of its shares into a public benefit fund, and its prospectus warns that its technology could pose catastrophic or existential risks to humanity.

TechCrunch's Rebecca Bellan reported the filing and noted that the structure means voting power will not match the founders' economic ownership, so public shareholders will have little practical control over the company. Anthropic's filing with the SEC sets out the full financials and governance. TechCrunch's earlier review of a leaked draft drew attention to the warning about existential risk.

The revenue growth is remarkable by any standard: more than 13 times in a year. But the more interesting part for anyone thinking about investing is what a share would and would not buy. Founders keeping majority voting control is common among big tech listings, and Anthropic's reason is unusual: protecting a safety mission. Still, the result is that public investors would put money into a company whose own filing says its technology could be dangerous, with little ability to influence how it is run.

It is the same tension we covered when Pepkor's shareholders got no vote on a major deal, at a far larger scale. The timing also puts last week's public benefit fund pledge in a new light, as one IPO adviser noted it helped with image ahead of the listing. The filing does not settle whether this is a sound structure or a protective one for founders, and reasonable investors will disagree.

What it does show is that the companies building the most powerful AI are now asking the public to fund them, on terms that keep the public out of key decisions.

You might also like our piece on Anthropic's AI safety bill, Claude becoming an office suite, and Meta's Muse Wonder.

Get more SA tech and business news and subscribe to The Open Letter.