Highlights
- Anthropic publicly filed its Form S-1 with the SEC on Monday, October 26, 2026, applying to list its Class A common stock on the Nasdaq Global Select Market under the ticker “ANT”, with Goldman Sachs and Morgan Stanley as joint lead underwriters
- First-half 2026 revenue was $16.286 billion, up about 1,300% from $1.163 billion a year earlier, while the net loss widened to $80.748 billion — most of it a roughly $72.4 billion non-cash charge tied to the rising fair value of its convertible notes
- The operating picture moved the other way: the operating loss narrowed to $693 million from $2.135 billion a year earlier
- The filing discloses computing commitments of more than $475 billion over the next decade, and its risk factors warn that advanced AI systems could pose “catastrophic or existential risks to humanity” and exhibit behaviour “resembling blackmail”
The filing: what is confirmed, and what is still blank
Anthropic, the developer of the Claude family of AI models, has filed for an initial public offering, announcing on Monday that it plans to list its Class A common stock on the Nasdaq Global Select Market under the symbol “ANT”, with the number of shares and the price range not yet set (Decrypt). Goldman Sachs and Morgan Stanley were listed as the joint leads on the deal (The Times), named first among 24 underwriters in total (Decrypt).
The public filing follows a confidential draft submitted to the SEC in June (NDTV Profit). In the document, Anthropic said the share sale would occur “as soon as practicable after this registration statement becomes effective” (The Times). Formal marketing of the offering could begin as early as the week of November 9, Bloomberg News has reported (Moneycontrol).
What the S-1 does not contain matters as much as what it does: there is no share count, no price range and therefore no official valuation. Every valuation number in circulation is a press estimate, not a company figure — a distinction the roadshow coverage will keep blurring, and readers should not.
The revenue curve: $1.16 billion to $16.3 billion in a year
Anthropic reported revenue of $16.286 billion for the six months ended June 30, 2026, compared with $1.163 billion in the year-ago period, representing growth of approximately 1,300% (NDTV Profit). The longer arc is just as steep: full-year revenue was $4.59 billion in 2025, up from $386 million in 2024 and $41 million in 2023 (Decrypt). In May, Anthropic raised $65 billion from investors at a $965 billion valuation (Investor’s Business Daily), and the company has said its annualised run-rate revenue crossed about $47 billion in May 2026 — a company-defined snapshot figure, not audited GAAP revenue, and one that should not be stacked directly against the $16.3 billion half-year result (LongYield).
Anthropic’s own claim for Claude in the filing is sweeping: it said Claude had experienced the “fastest enterprise adoption of any technology in history” (The Times). The concentration behind that adoption is a risk the filing states plainly. The company warned that “a substantial portion of our revenue is driven by a limited number of customers”, and said: “We anticipate that we will continue to derive a significant portion of our revenue from a limited number of customers for the foreseeable future” (The Times). Two customers reportedly accounted for about one-fifth of first-half revenue (NDTV Profit), and usage-based sales accounted for about 80% of first-half revenue (NDTV Profit).
The $80.7 billion loss that mostly is not one — and the compute bill that is
The headline loss needs unpacking, because two very different numbers sit inside it. Anthropic reported a net loss of $80.748 billion for the first half, and most of that is explained by a non-cash accounting charge of about $72.4 billion, primarily associated with changes in the fair value of convertible promissory notes as Anthropic’s valuation increased (NDTV Profit). In plain terms: the more valuable the company became on paper, the larger this paper loss grew. No cash left the business for that charge.
Underneath it, the operating loss narrowed to $693 million from $2.135 billion a year earlier (NDTV Profit), and operating cash flow was positive at $3.93 billion, although the company attributed part of that improvement to the timing of customer receipts and vendor payments (NDTV Profit).
The real financial weight is forward-looking. Computing and infrastructure expenses reached $8.945 billion in the first half, compared with $2.33 billion a year earlier, accounting for more than half of first-half revenue (NDTV Profit). The S-1 shows computing commitments of more than $475 billion for infrastructure over the next decade, including more than $100 billion to Google parent Alphabet, Amazon and SpaceX (Investor’s Business Daily). Broader reported totals run higher — existing infrastructure commitments are reported to exceed $600 billion, with additional deals under negotiation potentially taking the figure to approximately $750 billion, a total that covers cloud services and equipment arrangements and should not be read as $750 billion in booked cloud liabilities (NDTV Profit). The mismatch is the point investors will probe: long-term infrastructure agreements can require payments even when capacity is not fully used, while most of the revenue paying for them is usage-based (NDTV Profit).
This is not Anthropic’s first collision between safety positioning and commercial scale — we tracked that tension when the company backed California’s AI safety bill, and in a very different register when Moonshot apologised and reset quotas for Kimi K3.1 Pro under demand pressure.
The risk factors are the story: “catastrophic or existential risks to humanity”
Almost a third of the roughly 240,000-word document was dedicated to risk factors, including its AI models’ potential to “resist shutdown”, to “conceal or manipulate information” and to exhibit behaviour “resembling blackmail” (The Times). The filing’s own central warning, quoted in full, is this: “sufficiently advanced AI systems could, if not developed, deployed and operated with appropriate safeguards, oversights, and security measures, pose catastrophic or existential risks to humanity, particularly if such systems were misused, replicated, weaponised, or allowed to operate with a high degree of autonomy” (The Times).
Investor’s Business Daily carries the same passage from the filing: both Anthropic and outside researchers fear that “sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity”, particularly if such systems were “misused, replicated, weaponized, or allowed to operate with a high degree of autonomy” (Investor’s Business Daily). The Times also quotes the filing stating: “Losing control of sufficiently powerful AI systems could be a global security threat of the first order” (The Times). And on the direction of travel as models improve, the filing warns that the development of more advanced versions of the technology “could further increase the risk that our models cause harm” (The Times).
Set against that, in a letter signed by chief executive Dario Amodei and the founding team, the S-1 describes AI as “one of the most consequential technologies in human history” which, if managed well, could “accelerate scientific innovation” and greatly increase economic productivity — while also carrying risks from misuse, loss of control, authoritarian abuse and economic disruption (Investor’s Business Daily).
Control, valuation talk, and what happens next
Public shareholders will not control this company. Amodei and the other founders will hold 50.1% voting power through a Class F common stock, and the company’s Long-Term Benefit Trust will have power to appoint a majority of Anthropic’s board (Investor’s Business Daily). The filing warns that the structure “may effectively eliminate the ability of holders of our Class A common stock to influence important decisions, including a change in control” (Decrypt). Anthropic is incorporated as a public benefit corporation, which requires its directors to balance shareholders’ financial interests with its stated purpose: “the responsible development and maintenance of advanced AI for the long-term benefit of humanity” (Decrypt).
On valuation, treat every figure as an estimate attributed to the press, because that is what it is. Investor’s Business Daily frames the offering as one that could value the company at $2 trillion (Investor’s Business Daily). The Times reports a drift in the other direction: Anthropic is currently valued at $965 billion, was said to be targeting a $2 trillion valuation a few weeks ago, but is now thought to be targeting closer to $1.75 trillion — the valuation SpaceX achieved in its listing this year, the largest in history (The Times). Reporting on the earlier confidential draft had put backers’ hopes above $2 trillion (Decrypt). The $1.5–2 trillion range now in circulation is therefore best read as the outer band of press estimates, with the target reportedly settling nearer $1.75 trillion — the amended filing with a real price range, expected before any roadshow, is what will settle it.
How the market reacted
The first market response was positive, if hardly euphoric. The tech-heavy Nasdaq gained 0.92 per cent to 26,194.26 after the filing and the S&P 500 increased 0.70 per cent to 7,673.67, though both indices had faced selling pressure in recent weeks as investors worried about the sustainability of the AI boom (The Times). That reaction captures the IPO’s central tension in a single session: the same filing that produced a growth story also handed investors a $475 billion-plus spending commitment and a risk section unlike any in market history. The roadshow, whenever it begins, will be where those two readings of the same document finally get priced.