Anthropic Heads to the Stock Market at Up to a $1.4 Trillion Valuation Instead of the Expected $2 Trillion

The maker of Claude is preparing to go to Wall Street, but not at the price tag that was reported just recently. If you build a product on Anthropic's models or simply follow where the AI market is heading, this filing deserves a careful read. It's not about stock market drama, but about how investors are now counting money in the industry.

What the filing actually says

According to the document Anthropic filed with the US regulator on Tuesday, the company is selling 825 million shares at a price of up to $78. That yields up to $64.4 billion in proceeds from the offering and a valuation of up to $1.4 trillion. Separately, the company is raising another $26 billion through a private placement among anchor investors. They include Google's parent company, Alphabet, and chipmaker Nvidia. Trading could start as soon as next week.

Previously, many outlets reported that Anthropic was aiming for $2 trillion. Now the figure is noticeably more modest. For comparison: the largest IPO in history remains Elon Musk's SpaceX. The company went public at a valuation of $1.78 trillion and is now worth slightly less than $2 trillion.

A separate line in the updated filing is revenue. On an annualized basis it stands at $81 billion as of October 31, $10 billion more than a month earlier. If you roughly divide the upper valuation by that figure, you get about 17 years of annual revenue. This is our arithmetic, not data from the document, but it shows fairly well what growth the market is pricing in ahead of time.

Why the bar was lowered

The filing gives no official reason, but there is an analyst's opinion. Brian Hopkins of Forrester believes investors had doubts about the financial sustainability of the AI build-out even before. The results of the US midterm elections sharpened those doubts: opponents of data centers won. According to him, the market needed a reason to adjust valuations slightly, and the elections became that reason.

On top of this come old questions: how quickly, and whether at all, the gigantic infrastructure spending pays off. Add to that the espionage scandal with a Chinese trail around Nvidia. It hasn't hit the market as a whole yet, but, according to the Guardian, it has already affected Nvidia's own shares.

Note the paradox. Anthropic's revenue grows by $10 billion a month, and the valuation is still being cut. This means the market is no longer arguing about whether people are buying AI, but about the price of computing capacity and the political risks around it. Data centers need land, energy and permits, and local authorities and voters increasingly say "no."

What practitioners should do about it

First, a public company is obliged to regularly disclose its finances and risks. For teams that have tied their product to Claude, this is a source of data on supplier stability that didn't exist before. Read the reports the way you read an SLA.

Second, $26 billion from Alphabet and Nvidia means the bundle of "model developer, cloud and chip maker" is getting even tighter. For a startup, this is an argument to build the ability to switch between providers into the architecture. Lock-in to a single model and a single cloud is a separate cause for concern.

Third, data center politics now affects valuations and, in the longer term, possibly the availability of capacity. If you are planning a product with a heavy inference load, keep this on your list of risks.

Limitations

For now this is a filing, not a completed deal. The price is stated as "up to," and the final terms may differ. The reasons for the lower valuation are analysts' interpretation, not a statement from the company. And this is not investment advice: for a developer and a product manager, the main thing is to soberly assess the supplier's risks, not to guess the share price.