Anthropic is expected to pitch investors a $30tn market opportunity

Anthropic is likely to tell IPO investors that its potential revenue opportunity runs above $30tn.

That figure would top the $28.5tn SpaceX put in front of investors before its own record listing. Corrie Driebusch reported the number for The Wall Street Journal, citing people familiar with the matter.

Companies going public estimate a total addressable market to show investors how much room they still have to grow. The figure describes the annual revenue a company could theoretically capture at 100% market share. Bankers build it from industry data and their own models.

What $30tn looks like next to a real number

The 191 technology companies in the S&P 1500 booked $2.4tn in revenue last year, according to FactSet data cited by the Journal.

Anthropic’s figure is more than twelve times that. It describes a market in which one company could eventually earn a dozen times what the entire listed American technology sector currently earns between them.

The company is not claiming it will collect the money. A TAM measures the size of the pool, not the swimmer. But the pool is the argument for the price.

Anthropic arrives at it by looking at the full scope of work AI models could complete, the same people told the Journal. That framing counts human labour rather than software spending. This is how the number gets so large.

The escalation happened in seven years

When Uber went public in 2019, it claimed a $6tn market opportunity. It got there by counting the mileage value of every personal car and public transport service on earth. Wall Street called that aggressive at the time.

WeWork cited $3tn in the offering it eventually abandoned. That company never made it to market at all, which is its own footnote on how much a market estimate actually proves about a business.

SpaceX raised the ceiling in May, calling its $28.5tn the “largest actionable” market in “human history” in its filing. Most of that figure, $26.5tn, came from AI rather than rockets.

The rocket company, in other words, mostly sold investors an AI story. Anthropic is about to sell the same story without the rockets attached.

Someone already said this had gone far enough

Aswath Damodaran teaches finance at New York University. Wall Street calls him the Dean of Valuation, and he assessed the SpaceX number before its June offering.

The AI portion was “reaching the end of what’s plausible and pushing beyond”, he told the Journal at the time. Anthropic’s figure would sit $1.5tn above it.

The record TAM did not hold the record stock up

This is the part worth holding on to. SpaceX shares jumped at the offering and then gave it back, and by June the rally was unwinding.

They fell below $105 in intraday trading in early August. They now trade around the $135 offering price.

An investor who bought the largest addressable market in human history has made nothing on it in three months. That is the most recent evidence available on what a TAM is worth.

Anthropic would be pitching a bigger number into a market that has already watched the last one fail to hold.

What the company actually earns

Anthropic more than doubled its revenue in the second quarter, to $11.6bn. That is real growth on a real base, and it is up more than fourteenfold in a year.

The company projects roughly $190bn to $200bn of revenue in 2028, Reuters has reported. Its IPO valuation rests on those forecasts.

Hitting the top of that forecast would put Anthropic at under 1% of the market it is describing. The gap between what a TAM measures and what a company collects is the whole argument.

Why the number is not only theatre

A large addressable market does work beyond impressing a book of investors. It justifies a valuation, and it justifies the infrastructure spending that valuation has to support.

It also guides product priorities and frames the rivalry with OpenAI and Google. And it anchors the growth story behind the listing, Juby Babu noted for Reuters.

On that reading, $30tn is less a forecast than a statement about which business Anthropic thinks it is in. Counting the work rather than the software is the claim.

None of this is official yet

Anthropic has not published the figure. The company is still discussing its plans and the numbers could change, the Journal’s sources said.

The company did not immediately respond when Reuters asked for comment. It has said nothing publicly about a $30tn market.

Anthropic will file its IPO financial-disclosure documents within weeks. That timing could put the listing in September or early October.

The rest of the offering would break records too

Anthropic could aim to raise as much as $100bn, against the $86bn SpaceX took. It is targeting a valuation near $2tn, above the $1.77tn SpaceX reached.

Against the multiples AI companies already fetch, that valuation looks conservative rather than stretched. Which is its own comment on the state of the market.

Europe has no comparable listing to measure against. A $100bn raise would run to roughly ten times the largest European listing in decades.

What a number this size is for

A total addressable market has always involved guesswork. The Journal calls these estimates especially squishy when the question is how quickly AI reshapes entire industries.

Every company in this sequence stretched the definition of its own market until the number stopped constraining anything. Uber counted every car journey. SpaceX counted human history. Anthropic is counting work.

The test arrives with the prospectus, and it is a narrow one. Does the $30tn appear in the filing at all, and in what words? A pitch made in private and a number a company must stand behind are different things.