"Anthropic IPO: The $2 Trillion Filing That Lets the Public Buy In But Not Take Over"

On Monday, Anthropic published the document that starts turning a private AI lab into a public company. Its S-1, filed with the US Securities and Exchange Commission on 26 October, applies to list its Class A shares on Nasdaq under the symbol "ANT". Share count and price range are unset, and both the company and existing shareholders will sell stock, with Goldman Sachs and Morgan Stanley named first among 24 underwriters. The prediction market moved before most people had finished reading: on Polymarket, "Anthropic IPO before 2027?" jumped from about 75% to 94.5% within 24 hours of the filing going public. The market is calling a year-end listing close to a formality. I am slightly less sure, and the reason for the gap is worth examining, because it is not about the numbers.

The numbers are extraordinary. Revenue for the six months ended 30 June 2026 was $16.29 billion, up from $1.16 billion a year earlier, after $4.59 billion in all of 2025. The net loss for the first half of 2026 was $80.7 billion, which sounds catastrophic until the filing explains it: about $72.4 billion is a non-cash charge for changes in the fair value of convertible notes and other instruments, driven mainly by the rise in Anthropic's own valuation. The operating loss was $693 million. As of 30 June, the company had raised $141.1 billion in total and held $112.4 billion in cash plus $19.3 billion in short-term investments, and by 30 September it counted more than 1,900 customers above $1 million in trailing-twelve-month revenue. It is not listing because it is running out of money.

The next stage is priced, in the reports around the filing, at a level that would make this one of the largest offerings ever attempted. Anthropic raised $65 billion in May at a $965 billion valuation, and Reuters reported last month that its backers hoped for a valuation above $2 trillion. Legal reporting puts the expected raise at between $60 billion and $100 billion, with a listing expected in mid-to-late November. Those figures are reported expectations, not company guidance, and that distinction matters for a December deadline. A registration statement is not an effective offering: the shares cannot be sold until the SEC declares it effective. Still, the confidential draft went to the SEC in June.

What interests me most is not the valuation but the governance, because it is the one part of this filing that is genuinely novel. Anthropic is a Delaware public benefit corporation, with the stated purpose of the responsible development of advanced AI for the long-term benefit of humanity. A Long-Term Benefit Trust, made up of five financially disinterested trustees, will elect a majority of the board through a special class of stock. Immediately before the offering, a single Class F share will be issued to an entity owned by the seven co-founders, carrying 50.1% of the voting power on most shareholder matters. Holders of the Class A shares sold to the public get one vote per share, but the filing warns that the structure "may effectively eliminate the ability of holders of our Class A common stock to influence important decisions, including a change in control." The filing says Anthropic will be a "controlled company" under Nasdaq rules. The co-founders have each pledged to give 80% of their personal equity to philanthropic causes.

So the question this offering asks the public is a strange one: will you fund, at a possibly $2 trillion valuation, a company whose control you are explicitly told you will not have? The usual bargain of a listing is that money buys a voice, however small. Here the voice is small by construction, and the trust is the substitute: five trustees are supposed to hold the balance between profit and purpose that shareholders normally hold through votes. It is an honest design in that nothing is hidden. Whether it is durable will be tested by the first real conflict between what the market wants and what the trustees think responsible AI development requires. Dissenting shareholders will have one remedy: to sell.

That is why I put my confidence below the market's, at 85% against its 94.5%. The gap is not doubt about readiness; it is the friction of a giant offering, an SEC effectiveness process, and a price negotiation at an unprecedented valuation. If the listing lands in mid-to-late November as reported, the year-end deadline is comfortable. If it slips into January, the market's near-certainty will have been the thing that was mispriced, not the company. Either way, the filing is already a landmark in what it discloses: earnings, burn, and who is actually in charge. The public can buy in. It just cannot take over. That distinction, more than any valuation, is what this IPO will be remembered for testing.