Anthropic seeks a valuation of up to $1.4 trillion in its IPO, below the expected $2 trillion

Anthropic, the developer of the Claude chatbot, is seeking a valuation of up to $1.4 trillion (£1.1 trillion) in its imminent stock market listing. That is lower than the $2 trillion figure widely touted before investors began questioning the durability of the AI boom. The Guardian reports this based on a filing submitted to the US regulator on Tuesday, November 10, 2026. According to the newspaper, trading is expected to begin next week.

According to the filing, the deal is structured as follows. Anthropic aims to raise up to $64.4 billion by selling 825 million shares at up to $78 apiece. It will also raise a further $26 billion through a private placement with anchor investors, including Alphabet (Google's parent company) and chipmaker Nvidia. If demand is strong, an over-allotment ("greenshoe") option could add another $9.7 billion. Altogether, the total would come to around $100 billion if the shares sold at the top of the range.

The $1.4 trillion valuation corresponds to the top of the price range, $70 to $78 per share, and is calculated on a fully diluted basis. That includes all existing common shares, as well as employee stock options.

The backdrop is one of investor nervousness. The Guardian argues that doubts about the AI boom crystallized last week, when the US midterm elections delivered wins for opponents of data centers. This adds to earlier concerns, such as how quickly companies can earn a return on their AI infrastructure spending, or whether they ever will. Brian Hopkins, principal analyst at Forrester Research, says the vote gave investors the opportunity to correct valuations that were already worrying them because of the financial sustainability of the AI expansion. In his words, "the market needed an excuse to correct a bit," and the "election scare" was what triggered it, resulting in the lower valuation for Anthropic.

For reference, the largest IPO in history is that of Elon Musk's SpaceX. It debuted at a valuation of $1.78 trillion, raised $75 billion and is now worth just under $2 trillion. Its assets include the rocket launch business and the social network X. Even so, the roughly $100 billion Anthropic could raise would exceed SpaceX's haul. The article also mentions a Chinese espionage scandal affecting Nvidia, one of the pillars for AI investors. According to the newspaper, it threatens to cast a shadow over the sector, although so far it has only hit Nvidia's shares.

As for the founders, CEO Dario Amodei will sell about $260 million in shares, as will each of his six co-founders. At the top of the range, each co-founder's stake would be worth more than $18 billion.

The updated prospectus shows Anthropic's annual revenue running at a rate of $81 billion as of October 31, $10 billion more than the previous month. Its main expense is data center computing capacity, key to developing and running its systems. According to the prospectus, it has compute purchase and lease commitments of more than $70 billion over the next 12 months and $450 billion through 2030, with a further $450 billion through 2045.

Amodei explained the risk of these commitments on a podcast with Dwarkesh Patel earlier this year: he could buy a trillion dollars of compute starting in late 2027, and if his revenue did not reach that trillion, or even $800 billion, "there's no force on Earth, no hedge on Earth" that would save him from bankruptcy. Finally, the Financial Times reported that Goldman Sachs, which is leading the deal alongside Morgan Stanley, expects Anthropic to generate $600 billion in annual revenue in 2030.

Overall, the deal remains enormous in absolute terms. But the cut from expectations shows that the market is starting to demand proof that revenue will be able to sustain the industry's gigantic infrastructure commitments.