Anthropic Just Filed for Its IPO — $16.3 Billion in Revenue, an $80.7 Billion Loss, and What It Means for Developers Building on Claude (October 2026)
I have read a lot of IPO filings, and most try to make a complicated business sound simple. Anthropic's filing, made public on October 26, 2026, does the opposite: it makes a simple story — a company growing very fast — sound complicated, because the numbers underneath it genuinely are. Revenue up more than thirteenfold in a year. A loss five times bigger than that revenue. A governance structure designed to stop shareholders doing what shareholders normally do. And a ticker symbol, ANT, for a company that has not yet said how many shares it will sell or at what price.
If you build on Claude, or you are thinking about it, this filing matters more than the headlines about valuation. The valuation reports are noise for now. The filing itself is the document that will govern how this company behaves once it is public, and it tells you a lot about what to expect from its API, its pricing, and its priorities.
What actually got filed
Anthropic, PBC filed a registration statement on Form S-1 with the United States Securities and Exchange Commission for a proposed offering of its Class A common stock. The company said it intends to list on the Nasdaq Global Select Market under the symbol ANT. The number of shares to be offered and the price range have not yet been determined, and the registration statement has not yet become effective — which means, as Anthropic itself noted, no shares can be sold and no offers to buy can be accepted until it does become effective.
That point is worth pausing on: a public S-1 is the start of a process, not the end of one. Anthropic had already filed confidentially in June 2026, according to coverage, so this public version is a step forward, not a first step. There is still a roadshow, a pricing decision, and an effective date to come. Some coverage has pointed to a possible listing in mid-to-late November, but treat that as reporting, not a date Anthropic has announced.
The underwriting syndicate is large. Goldman Sachs and Morgan Stanley are leading it, alongside J.P. Morgan, Citigroup, Barclays and Wells Fargo Securities — twenty other banks in total, according to coverage of the filing.
The revenue numbers, stated plainly
Here is the part that made the market sit up. According to the filing, Anthropic's revenue was $41 million in 2023, $386 million in 2024, and $4.6 billion in 2025. In the six months ended June 30, 2026, revenue reached $16.3 billion, compared with $1.16 billion in the same period a year earlier — a year-over-year growth rate the filing puts at 1,300 percent, and which other summaries calculate at 1,304 percent.
That curve is not normal. It is the kind of growth you get when a technology moves from interesting to infrastructural in a very short window, and when enterprises start buying it in bulk rather than experimenting with it. Anthropic has mainly focused on the enterprise market, and the filing reports net revenue retention of about 900 percent — how much existing customers increase their spending over time. In plain terms, the customers it already has are spending far more, not just renewing.
But Anthropic itself warns it does not expect revenue to keep growing at comparable rates, and cannot assure it will keep growing at all. A 1,300 percent growth rate is a historical fact about the last year, not a forecast — and Anthropic is explicitly telling you not to treat it as one. Coverage also reports five customers account for roughly a quarter of revenue, so if one renegotiates or slows down, it shows up in the next quarter's numbers.
Why the $80.7 billion loss is not quite what it looks like
Now the number that looks impossible. The filing shows a net loss of $41.97 billion for 2025 and $80.75 billion for the first six months of 2026. On revenue of $16.3 billion, a loss of $80.7 billion sounds like a company burning money at a rate that cannot be sustained — and in a sense it is, but the composition of that loss matters enormously, and it is the part most headlines skip.
According to summaries of the filing, $72.4 billion of the first-half loss was a non-cash change in the fair value of convertible notes, derivative liabilities and SAFEs — an accounting charge tied to the rising value of instruments that will convert into shares, not cash walking out of the door in that period. That does not make the loss meaningless: those instruments represent real future dilution. But you should not read $80.7 billion as cash spent on salaries, chips and data centres in six months.
The honest way to hold both facts at once is this: Anthropic is spending heavily, as every frontier AI company is, and it is also carrying a financing structure whose accounting swings wildly when its valuation moves. Going public should simplify that structure over time, because the convertible instruments convert and the fair-value swings go away.
The governance part developers should actually read
Anthropic is a public benefit corporation, founded in 2021 by seven former OpenAI employees, including chief executive Dario Amodei and president Daniela Amodei. I wrote earlier about Anthropic pledging a share of itself to a public benefit fund, and this filing shows how that commitment is wired into its structure.
The filing describes a Long-Term Benefit Trust, which it calls a financially disinterested trust, that generally has the power to elect and remove a majority of the board. The seven co-founders will hold a special Class F stock through a founders' entity. In practice, control is kept away from ordinary shareholders in a way unusual for a normal public company. The point, as Anthropic has always framed it, is to protect its mission — building safe, beneficial AI — from being overridden by short-term shareholder pressure.
Whether you think that is admirable or alarming probably depends on where you sit. As a developer, the practical question is narrower: does this structure make Claude's roadmap more stable, or less accountable? A board that cannot easily be removed is less likely to pivot wildly to chase a quarterly number — but also less likely to respond quickly if developers dislike a pricing or policy change. Both can be true at once.
The filing also flags risks in unusually blunt language: Anthropic warns its AI could behave in unintended, misaligned or autonomous ways, and could pose catastrophic or existential risks if misused, replicated, weaponised or allowed to operate with a high degree of autonomy. The prospectus mentions attempts to resist shutdown and to conceal or manipulate information. You do not normally see a company going public warn, in its own offering document, that its product could pose existential risks.
What this means if you build on Claude
First, expect more scrutiny, not less stability. A public Anthropic will report quarterly, explain its compute spending, and answer questions about margins and customer concentration. That scrutiny tends to push companies toward clearer pricing and more predictable product decisions, because surprises get punished by the market. For a developer, predictability is worth a lot. An API whose pricing changes without warning is a liability in client work.
Second, expect compute to stay at the centre of the story. Claude is available on Amazon Web Services, Google Cloud and Microsoft Azure, and Anthropic trains across Amazon Trainium, Google TPUs, Nvidia and AMD chips. That diversification is deliberate: no single chip supplier, no single cloud. For developers, the benefit is resilience; the cost is complexity, and complexity is where outages hide.
Third, do not build your whole business on a rumoured valuation. Reports have varied widely: one said Anthropic is targeting as much as $2 trillion and aiming to raise $100 billion, while other reporting put recent investor valuations closer to $1.5 trillion. The filing gives no valuation, no share count and no price range. A valuation tells you nothing about what Claude will cost you next year.
Fourth, take the enterprise focus seriously. Enterprise focus usually means better reliability, support and compliance tooling — but also product decisions optimised for large customers first. Keep an exit plan: abstract your prompts and tooling so switching providers is painful but possible, not a rewrite.
What could still change, and the traps to watch
A few things to watch between now and the first day of trading. The share count and price range will be disclosed in a later filing, and those will tell you far more about the deal's size than any current report. The roadshow will test whether institutional investors believe the growth story at the price Anthropic wants.
The traps are familiar, but they apply with unusual force here. Do not confuse revenue growth with profitability, or a non-cash accounting loss with a cash crisis. Do not trade on the rumoured valuation — until Anthropic names a price range, you are guessing. Do not assume the governance structure means the mission is safe forever; structures can be amended, and public-market pressure is patient. And do not ignore concentration risk.
Conclusion
Anthropic's S-1 is a strange and revealing document: a company that went from $41 million in revenue in 2023 to $16.3 billion in half a year in 2026, lost $80.7 billion on paper in that same half year, and is asking public investors to buy into a structure designed to limit their control — while warning, in its own words, that its technology could pose existential risks if mishandled.
My take, as someone who builds on these tools rather than trades them, is simpler than the headlines. Anthropic going public will make Claude more scrutinised, more predictable, and more exposed to market and regulatory pressure, all at once. That is mostly good news for developers who want a stable platform, as long as you do not mistake a filing for a promise. Keep your architecture portable, and judge Claude on what it does for your work — not on what the market says it is worth on day one.
FAQ
What did Anthropic actually file?
Anthropic, PBC filed a Form S-1 with the US SEC for a proposed offering of Class A common stock, intending to list on the Nasdaq Global Select Market as ANT. The filing is not yet effective, so no shares can be sold until it is.
How many shares will Anthropic sell, and at what price?
Not yet determined. The filing gives no share count or price range; both will come in a later filing. Reported valuations up to $2 trillion or around $1.5 trillion are unconfirmed reporting, not figures from the filing.
How fast is Anthropic actually growing?
Very fast: $41 million in 2023, $386 million in 2024, $4.6 billion in 2025, and $16.3 billion in the first half of 2026, up from $1.16 billion a year earlier. Anthropic cautions it does not expect comparable growth to continue.
Why did Anthropic lose $80.7 billion in six months?
The reported first-half 2026 net loss was $80.75 billion, but about $72.4 billion was a non-cash change in the fair value of convertible notes, derivative liabilities and SAFEs, per filing summaries. It is largely an accounting charge, not cash spent in the period.
What is the Long-Term Benefit Trust?
A financially disinterested trust described in the filing as generally able to elect and remove a majority of Anthropic's board. With the founders' special Class F stock, it is designed to keep Anthropic's mission from being overridden by ordinary shareholder pressure.
Should developers switch to, or away from, Claude because of the IPO?
Not on the IPO alone. It suggests more scrutiny and predictability ahead, but no immediate change to product or pricing. Build on Claude if it serves your work, keep your tooling portable, and wait for later pricing filings before big commitments.
