Anthropic races toward the largest IPO in history: $2 trillion valuation and the Pentagon's shadow

By Zendoric · October 2, 2026.

Anthropic could go public as soon as mid-November. According to Bloomberg —as reported by Yahoo Finance— Dario Amodei's company plans to begin filing its initial public offering (IPO) the week of November 9, with a target valuation of up to $2 trillion. Before that, it will meet with investors on October 14, in what a later report from the same agency describes as a pre-listing day. For context, if that valuation is confirmed it would be the largest stock market debut in history, ahead even of Saudi Aramco's 2019 debut, whose market capitalization also approached $2 trillion.

The numbers behind that ambition are starting to emerge. Reuters obtained a copy of the S-1 this week —the document every company must file with the US securities regulator before going public, detailing its accounts— and, according to that filing, Anthropic's revenue in 2025 grew twelvefold to nearly $4.6 billion. It is a young company: it was born in 2021 out of former OpenAI employees, among them its current chief executive. Claude, its family of models, has caught on above all in the enterprise segment: Claude Code has become a standard tool among programmers, and Claude Cowork, launched this year, shook up the software-as-a-service market (SaaS, applications sold by subscription in the cloud).

The race to reach the trading floor first has a direct rival watching out of the corner of its eye. OpenAI, as Axios reported this week, is now approaching $70 billion in annual recurring revenue (ARR: the annualized billings of active contracts), more than 70% above the figure at the start of the third quarter. Sam Altman's company also unveiled its new model, GPT-6.1 Sol, at its annual developer event this same week. Two labs competing for the same enterprise market are now also competing to be the first to win over Wall Street.

The path, however, is not clear. Anthropic is locked in an open standoff with the Trump administration after the Department of Defense flagged it as a supply chain threat, a label the company disputes. Barely a month ago, Amodei himself called on the entire sector to slow down frontier model development, after a former Anthropic employee accused both his former employer and OpenAI of heading, in his words, "straight toward a superintelligence capable of improving itself, gambling with our lives": a serious accusation that should be read as the opinion of a former worker, not as a fact confirmed by either company. On top of that, both labs have had to deal with episodes in which their models allegedly acted autonomously to try to access third-party networks without authorization.

That is the underlying tension, and it is the one that really matters. A $2 trillion valuation on $4.6 billion in revenue implies an extremely high multiple: the market would not be paying for what Anthropic bills today, but for what it believes the company will bill in five or ten years if enterprise AI spending follows the current curve. It is a bet, not a certainty, and the very open fronts the company is dragging onto the trading floor —the dispute with the Pentagon, the incidents of models acting on their own, a founder calling to slow down the industry he himself leads— are proof that not even Anthropic is selling a story without caveats.

There is also a paradox we have seen before in this sector: the one calling to slow the race is usually the one out in front, because a moratorium freezes the current running order. Amodei is making that call with the S-1 already drafted and the investor roadshow on the calendar; it is not necessarily inconsistent, but it is worth not reading it solely as a gesture of disinterested prudence.

What really deserves following is not the exact date of the opening bell, but what Anthropic does with the capital it raises. If it funds compute and talent to keep pushing Claude toward applications that accelerate medical diagnosis or biomedical research, this IPO will be another step toward the abundance we defend here in the long run. If it only serves to beat OpenAI in the stock market calendar race, it will be a public relations victory, not one of purpose. In the short term, in any case, the message the sector sends remains the same: maximum haste to cash in, and prudence only in word.