For a few strange hours on Monday evening, the most important document in artificial intelligence did not officially exist. Bankers had been sent the pricing terms for the Anthropic IPO — and then, just as quietly, the company took them back.
By Tuesday, the real thing was public. According to Reuters, Anthropic had filed an updated prospectus for its stock market debut, and the numbers inside were big enough to explain all the drama.
This is the moment the AI industry has been circling for years: a frontier AI lab asking the public markets to put a price on the future it keeps promising. The answer Anthropic is proposing is about one trillion dollars.
Inside the Anthropic IPO Filing: 825 Million Shares at $70 to $78
The company plans to sell 825 million Class A shares. Roughly 801 million of them are new shares issued by Anthropic itself, while about 23.6 million will be sold by existing shareholders cashing out part of their stakes.
Anthropic is explicit that it will receive nothing from those secondary sales. Every dollar that funds its future comes from the new shares — and at the top of the range, that primary sale alone is worth about $62.5 billion.
Count the shares sold by existing holders, and the public offering could raise as much as $64.4 billion in total. Underwriters also get a 30-day option, known as a greenshoe, to buy roughly 123.7 million additional shares.
The shares are set to trade on the Nasdaq under the ticker symbol ANT. Goldman Sachs and Morgan Stanley are leading the syndicate of banks running the deal, according to the filing.
A $1 Trillion Price Tag — Below the Hype, Above the Last Round
Based on the share count in the prospectus, Reuters calculates that pricing at $78 would value Anthropic at roughly $1.02 trillion. Add in options and unvested stock, and the fully diluted figure climbs toward $1.3 trillion.
By any normal standard, that is a staggering valuation. Yet inside the AI boom’s strange economics, it is also a step down: investors had at one point talked about $1.5 trillion, and earlier this year even $2 trillion.
It still represents a step up from where Anthropic last marked itself. A private funding round in May valued the company at $965 billion, so the IPO target is ambitious without being detached from its own recent history.
The timing, though, is delicate. The Nasdaq recently suffered its deepest one-day drop in 19 months as investors questioned AI valuations, before rebounding sharply on Monday — a reminder of how quickly sentiment around this sector can swing.
There is a giant shadow over the comparison table, too. In June, Elon Musk’s SpaceX raised $75 billion in the largest IPO in history, at a valuation near $1.75 trillion. Anthropic’s debut would be the AI industry’s answer.
Tech Giants Are Adding Another $26 Billion on Top
The public offering is only part of the money. Sitting alongside it is a separate $26 billion private placement, in which some of the biggest names in technology have agreed to buy Anthropic’s Class N stock after the IPO closes.
Nvidia and Alphabet’s Google lead the cornerstone list at $10 billion each. AMD is buying $2.5 billion, Microsoft, Broadcom and Micron are in for $1 billion apiece, and TSMC rounds out the group with $500 million.
Altogether, the public raise plus the placement could bring Anthropic’s total fundraising to about $90.4 billion — one of the largest capital events in corporate history, assembled around a laboratory founded only in 2021.
The structure says a lot about how the AI economy now works. The companies selling Anthropic its chips and cloud capacity are becoming its shareholders too — supplier, partner and owner, all in the same corporate bodies.
The Numbers Behind the Anthropic IPO: A $17 Billion Quarter, $900 Billion in Bills
The prospectus also opens a window into Anthropic’s finances. Revenue reached $16.96 billion in the most recent quarter, up 47% from the quarter before it, while the company’s net loss narrowed over the same period.
Growth on that scale is the engine of the trillion-dollar ask. A business adding billions in quarterly revenue can plausibly argue that today’s eye-watering valuation is simply tomorrow’s normal, marked early.
But the same filing describes close to $900 billion in compute and infrastructure commitments stretching years into the future, owed largely to cloud partners such as Amazon and Google. The growth machine runs on borrowed capacity.
That is the central tension of this listing. Anthropic is expanding at a pace almost no company has ever matched, while signing bills almost no company has ever owed. Both facts are true at once.
None of it has slowed the product race. This week the company launched its newest flagship models — as we reported in our coverage of Claude Fable 5.2, they are its fastest yet, and cheaper to run than what came before.
Why This Listing Is a Referendum on the Whole AI Boom
Strategists are treating the debut as more than one company’s payday. UniCredit’s Christian Stocker called it a test of how convinced investors are about the pace of this technological transformation — and warned a flop could reprice the entire AI story.
There is simply no template for what happens next week. No frontier AI lab has ever faced public-market pricing before, so every assumption — about growth, about losses, about the cost of compute — gets tested in the open for the first time.
Anthropic’s own filing captures the strangeness. According to CNN, the prospectus promises technology that could “bring forth an age of miracles” while warning, in the same breath, that the technology might one day endanger humanity itself.
The company is also arriving on the public stage with political visibility to match. Days ago it stood among the ten AI firms that signed the Paris Call for AI Safety and Security, a pact meant to put guardrails around frontier models.
Anthropic declined to comment on the filing when reporters asked. In a week like this, the prospectus — hundreds of pages of numbers, risks and prophecy — is doing all the talking on its behalf.
What Happens Next: The Roadshow, Then a Debut as ANT
The roadshow began on Tuesday, with Anthropic pitching investors through virtual and in-person meetings. According to the Wall Street Journal, the company aims for its shares to start trading on Thursday or Friday of next week.
The final price will only be set once those meetings read the market’s mood. Monday’s retracted draft was proof of how fluid the terms still are — a deal this size can be repriced until the last evening.
Meanwhile, the infrastructure race funding all of this keeps accelerating. Google unveiled new TPU systems this week to feed the same demand for compute that fills Anthropic’s commitment sheets — read our breakdown of the TPU 8tx and 8i Max launch.
Within days, then, the experiment begins. A company that sells intelligence by the token will watch its own value flicker on a public ticker — and the whole AI industry will learn what its promises are actually worth.
