Anthropic reportedly targets a $2 trillion-plus IPO with a launch the week of Nov. 9
Bloomberg reported on Oct. 1 that Anthropic hopes to begin the formal marketing of its IPO during the week that Nov. 9 falls, allowing shares to trade before Thanksgiving. The prospectus Reuters has seen shows a net loss of more than $42 billion in 2025, planned infrastructure spending of $518 billion — and unusually candid warnings about "existential risks to humanity."
What's new
According to SiliconANGLE, citing Bloomberg, Anthropic has set a new target of beginning to formally market its IPO during the week that Nov. 9 falls. That means shares would begin trading before Thanksgiving, and after the Nov. 3 midterm elections. The sources are anonymous, and SiliconANGLE notes that internal discussions about the timeline are still ongoing and could change. The company is, according to the reporting, determined to be listed before the end of the year.
The date updates what Reuters had reported earlier in the same week: that the company would most likely wait until just after the midterms. With OpenAI's listing apparently delayed, Anthropic's timeline has become one of the most important calendar entries for AI investors. SiliconANGLE itself points out that a forced delay from Anthropic, like the one OpenAI already appears to have carried out, could create a shock that damages other AI stocks.
The listing that would be the largest in history
The process began in June, when Anthropic filed a confidential S-1 document with the SEC, The Motley Fool reports. An October listing was the original expectation before it was recently pushed to November. The company is set to seek to raise up to $100 billion, at an expected valuation above $2 trillion. That would make Anthropic, best known for the Claude chatbot, one of the world's most valuable companies — and the listing the largest in history, ahead of SpaceX's June listing at $1.77 trillion.
The numbers in the prospectus
Reuters' journalists have seen the prospectus. According to SiliconANGLE and The Motley Fool, it shows:
- Fiscal 2025 revenue of $4.6 billion, growth of more than 12x from the year before
- Operating losses of more than $8 billion
- Net loss of more than $42 billion
- Planned spending on cloud, computing power and infrastructure in coming years of $518 billion
- Around 80 pages devoted to risk discussions
The large difference between the operating loss and the net loss is not explained in the available coverage.
Unusually candid risk warnings
Among the 80 risk pages, several are dedicated to what the prospectus calls "existential risks to humanity," SiliconANGLE reports. According to the coverage, the warnings address, among other things, models showing "self-preserving behaviors," attempts to "conceal or manipulate information," and behavior resembling blackmail — without further details being given.
The coverage itself points to the tension with the company's own rhetoric: In September, CEO Dario Amodei publicly urged AI model developers to "pace the frontier" — a coordinated industry effort to slow the development of frontier models so the technology can be advanced safely. Amodei has also proposed placing third-party observers in the development teams. The tension between the prospectus's warnings and the rush toward listing is not commented on by the company in the available coverage.
The financial picture: two versions of Anthropic
The prospectus's 2025 figures stand in contrast to the separately reported figures for 2026. Anthropic itself has stated that the company's annual run-rate (ARR) started 2025 at $1 billion, rose to $9 billion at the end of the year, reached $47 billion in May and stood at $65 billion at the end of July 2026 — above the reported ARR figure of $40 billion for OpenAI. Bloomberg has, according to The Motley Fool, reported Q2 revenue of $11.5 billion, and the Financial Times reported in September that Anthropic was on course to post an adjusted operating profit for the second straight quarter, with gross margins of 80 percent before revenue shared with distribution partners.
It is worth keeping two things apart here. First, ARR and recognized revenue are not directly comparable — ARR is an annualization of ongoing sales, not booked revenue. Second, the prospectus, as described, contained no financing figures for 2026. The reported 2026 strength and the prospectus's 2025 figures can both be true and still tell two different stories — the gap is not closed in the available coverage.
The spending behind the $518 billion
Part of the planned infrastructure spend can be read directly from the company's own contracts. In May, Anthropic announced it will pay SpaceX $1.25 billion per month through May 2029 for use of the Colossus 1 data center — $15 billion per year. In August it entered a six-year cloud deal with Nscale worth $44.6 billion. These two contracts alone make up a significant share of the commitments, but the prospectus reporting does not specify the full breakdown of the $518 billion.
What is confirmed — and what is not
None of the central claims are verified against primary sources in what is available. The timing in the week around Nov. 9 comes from Bloomberg's anonymous sources, via SiliconANGLE, and the timeline can still change, according to the reporting. The prospectus figures and the risk language are known through Reuters' reading of the document, relayed by SiliconANGLE and The Motley Fool. There is no public S-1 filing and no primary statement from Anthropic confirming either the date, the valuation or the financing figures. OpenAI's delay is also only third-party reported.
What the S-1 actually says — and whether the listing happens before Thanksgiving — remains something the market must wait for the filing itself to answer.