Anthropic, the San Francisco artificial intelligence developer founded by former OpenAI researchers, prices its IPO on Wednesday at $70 to $78 per share and begins trading under the ticker ANT on Nasdaq on Thursday, seeking up to $64.4 billion in a listing that could value the company at $1.02 trillion — the second-largest IPO in history, behind SpaceX, according to Novinky.
That headline number, reported by Novinky, is not the whole picture. According to E15.cz, the primary offering plus private placements could push past the record set by SpaceX’s June listing and make this the largest IPO ever. The two outlets frame the same deal differently on the raise; both agree on the upper valuation: at $78 a share, Anthropic crosses the trillion-dollar mark.
This is the math investors are actually wrestling with. The prospectus, as covered by iROZHLAS, shows revenue growing fast but no profit — the last quarterly loss exceeded $23 billion — alongside a commitment to spend close to a trillion dollars on computing capacity in the coming years. E15.cz puts the contracted figure for chip and server rental alone at $417 billion.
“From this point of view, the timing of the IPO is not entirely ideal. Investors have become significantly more sensitive to the return on AI investments in recent weeks,” said Marek Malina, analyst at Portu.
What Thursday means for your order ticket
For retail investors, the mechanics matter as much as the valuation. Czech investors were excluded from the primary offering — retail allocations on European platforms covered only selected Western European and Nordic countries, Novinky reported.
The way in is the open market: once the opening auction on Nasdaq closes on Thursday evening, likely between 5:30 p.m. and 7:00 p.m. Central European Time, ANT becomes tradable through any broker. Portu analyst Marek Malina flagged the timing on Wednesday.
“Retail allocation through European platforms is limited to selected Western European and Nordic countries, and Czechia is missing from it. Investors can get involved only from Thursday evening, when the stock starts trading publicly under the ticker ANT and becomes available through any broker,” said Marek Malina, analyst at Portu.
Malina also pointed to the passive route: index and technology ETFs. Anthropic qualifies for an accelerated entry into the Nasdaq 100 almost immediately, while the S&P 500 is at least a year away.
Invezz’s index calendar fills in the dates: MSCI could add ANT effective around November 30 or December 1, with FTSE Russell additions on December 14 and December 21, and the Nasdaq 100’s December 21 reconstitution seen as a possible fast-track date. S&P Dow Jones Indices has refused to waive its seasoning and profitability tests for megacap companies, per Invezz, so that door stays shut for now.
The circular money problem
But the shareholder register tells a stranger story. Roughly a quarter of the shares in the offering are being taken by Anthropic’s own competitors and suppliers — Google, Microsoft, Nvidia and other chip and infrastructure makers, iROZHLAS reported.
That creates companies financing the capital of the firm that buys their chips, meaning part of the demand for AI infrastructure is funded by the very companies that profit from building it. If demand for Anthropic’s services does not grow fast enough, iROZHLAS notes, that self-financing loop can turn financially fragile.
The contrast with the supplier side is stark. Nvidia, whose results E15.cz covered this week, more than doubled its net profit year over year to $65.3 billion in its third fiscal quarter, with revenue up 103% to $115.8 billion — its eighteenth consecutive beat of Wall Street expectations. Anthropic’s $23 billion quarterly loss sits on the other side of that ledger. One company sells the shovels; the other is digging, on credit.
| Metric | Figure | Source |
|---|---|---|
| Price range | $70–78 per share | Novinky |
| Primary offering target | up to $64.4 billion | Novinky |
| Valuation at top of range | ~$1.02 trillion | Novinky / E15.cz |
| Last quarterly loss | over $23 billion | iROZHLAS |
| Committed chip/server rental | $417 billion | E15.cz |
A public-benefit company with uninsurable risks
There is a governance wrinkle, too. Anthropic is chartered as a public benefit corporation, committed in its own words to the responsible development of advanced AI for the long-term benefit of humanity — a mandate that sits awkwardly with investor appetite for dividends, iROZHLAS notes.
Its prospectus admits its top risks are uninsurable: the company could face significant and unpredictable legal claims if its AI were misused in a catastrophe insurers could not cover.
And the founders’ own risk assessments are on the record. iROZHLAS points out that research lead Evan Hubinger has warned of a meaningful probability of AI wiping out humanity within ten years, while CEO Dario Amodei sees a real risk of catastrophic outcomes from AI development. Neither assessment has stopped either man from accelerating the work.
What the market is really pricing
As Malina framed it in Novinky, the listing is a test of how high the market is willing to value the business of building frontier AI models itself. The prospectus shows revenue growing significantly faster than computing-infrastructure costs — the bull case in one line.
The bear case is that companies like Anthropic must prove, in practice, that today’s enormous spending on chips and data centers produces an economic return, and recent weeks have made investors noticeably less patient on that question.
So Thursday evening is a referendum with three possible readings: a trillion-dollar endorsement of frontier-model economics, a warning shot if the stock trades below its $70 floor, or something in between once index funds start their rule-driven buying in December. The company has priced its ambition at $78. What it has not yet answered — and what the prospectus cannot answer — is when $23 billion quarterly losses turn into the profits the valuation assumes.