Anthropic’s $2trn IPO is priced below what AI stocks already fetch

The number comes from investors, not the company. Six backers told the Financial Times that rising revenue would let the five-year-old lab more than double its valuation in an autumn listing.

At $2 trillion it would eclipse SpaceX, which went public at $1.77 trillion in June.

Anthropic itself has fixed nothing. Several investors said senior executives had not settled on a target, even privately. The models are the investors’ own.

The machinery is moving regardless. Morgan Stanley, Goldman Sachs and JPMorgan are leading the offering, Quartz reported.

Three bulge-bracket banks on a mandate is not a rumour. It is the difference between investors discussing a listing and a listing being built.

The revenue, and the caveat attached to it

Two figures matter. In May the company said its annualised revenue had passed $47bn, which is the last number Anthropic disclosed itself.

Investors now expect that to reach $100bn to $120bn by the end of 2026, a rise of more than ten times across the year.

Read the definition before the number. Annualised revenue is Anthropic’s preferred measure. It infers full-year sales from recent performance rather than counting a year of receipts.

That is a legitimate way for a fast-growing company to describe itself. It also means a strong month flatters the figure. It is not the number a public company reports.

The multiple is less mad than it sounds

Run the arithmetic on the disclosed figure. At $2 trillion against $47bn, Anthropic would trade at roughly 43 times revenue.

Anthropic has no listed American peer to price against. But companies treated as AI beneficiaries have traded this year at about 55 times revenue, the FT notes, naming Palantir and the cloud group Nebius.

On that comparison, $2 trillion is not an outlier. It is below where the market already values businesses with a fraction of Anthropic’s growth rate.

One of the investors went further. “If Anthropic is growing 800 percent a year, you’d think at the incredibly low end they would trade at 30 times,” the person said. “That would make them a $3 trillion company.”

The first risk is the American government

Here is what the multiple does not price. Anthropic is in active litigation against the Department of Defense, which labelled it a supply-chain risk.

In June the Commerce Department hit the company with export controls, and Anthropic briefly pulled Fable 5 and Mythos 5, its two leading models.

That episode has a number attached. Two investors told the FT the temporary ban slowed revenue growth in June, before the company rebounded.

So a regulator bent the growth curve once already this year, inside a single month. The episode also spooked customers who had built on those models.

The second risk is that customers are looking at the bill

Anthropic’s market-leading model costs more than two and a half times as much to use as OpenAI’s flagship, according to Artificial Analysis. Chinese open-weight models cost a fraction of either, and they improved sharply this year.

The company gained market share among US businesses last month, on payments data from Ramp. The same analysts found those businesses were “hitting their limit on AI spend”. Many are moving to cheaper alternatives.

Some have gone further than switching models. Faced with rising costs, certain customers have reversed internal instructions telling employees to maximise their AI use.

That is the uncomfortable pairing. Anthropic is winning share in a market whose buyers want to spend less, while Chinese models keep closing in on capability.

Why the company is saying nothing

Anthropic declined to comment, and it has a reason beyond preference. It filed confidentially with the Securities and Exchange Commission in June. That puts it in a quiet period, limiting what it can say about its finances.

So every number in circulation comes from investors who stand to gain from the listing. That is not a reason to dismiss them. It is a reason to attribute them carefully.

The money behind those models is not small. Venture capitalists, sovereign wealth funds and other institutions have put just under $100bn into the company during 2026 alone.

Its valuation passed OpenAI’s for the first time in May, at $965bn post-money, and secondary trades have since marked it at $1.2 trillion.

What the bulls actually say

The case for is worth stating in its own words. Anthropic has spent the year releasing models that outperformed rivals, and selling them to businesses rather than consumers.

One investor, who has also backed OpenAI and SpaceX, put it plainly to the FT. “It’s easy to come up with challenges,” the person said. The company “continues to be in first position in performance, positioning, and what people want exposure to”.

There is truth in that last clause. A listing this size is partly a bet on investor appetite for AI exposure, and appetite is not the same thing as fundamentals.

Which is why the FT frames the risk as a market one. A float at $2 trillion would test public markets that are growing more nervous about the AI boom.

What would settle it

Two things, and the first is close. A public S-1 replaces every investor model with audited revenue. It will show whether the $47bn disclosed in May grew the way the projections assume.

The second is the June dip. If one month of export controls measurably slowed growth, the filing will show it. That line matters more than the headline number.

Until then the honest summary is narrow. The multiple is defensible on the numbers available. But the company cannot confirm them, its customers are trying to spend less, and its own government has already switched its best products off once.