Anthropic’s S-1 shows $16.3 billion of revenue in the six months to June 2026, up from $1.2 billion in the same period a year earlier (Anthropic S-1, 2026). It also shows an $80.7 billion net loss. Both numbers are true, and the second one comes from accounting, not from cash running out.
I use Claude every day, for coding and for workflows that need to give the same result every time. So when the company filed to go public on 26 October, I went to the filing, called an S-1. That is the document a US company must file with the regulator before it lists its shares. I read it for the risk factors, because those describe the supplier I depend on, and skipped the share price.
This article is for owners and managers who build on Claude. It is not investment advice, and it does not tell you to buy or avoid anything. It covers what the filing says and the risks it exposes for a customer. I last checked it against the S-1 on 28 October 2026.
Key Takeaways
- Anthropic filed a public S-1 on 26 October 2026 to list Class A shares on Nasdaq as ANT. No price or date is set.
- The $80.7B first-half net loss sits beside $81.1B of paper revaluation charges. On Anthropic’s own adjusted measure, it made $2.7B.
- For a customer, the IPO does not add risk. It makes three risks readable: big customers, fixed compute bills, and government decisions.
- In June 2026 the US Department of Commerce restricted two Anthropic models and the company switched them off for every customer until the restrictions were lifted.
What Did Anthropic File, and When Does It List?
Anthropic filed a public S-1 on 26 October 2026 to list Class A shares on the Nasdaq Global Select Market under the ticker ANT.
The filing says the number of shares and the price range “have not yet been determined” (Anthropic, 2026). It also says the registration statement “has not yet become effective”, so nobody can buy shares yet. Press reports expect a listing in November, but they disagree on the week, so treat the date as unconfirmed.
| Question | What the filing says |
|---|---|
| Where will it list? | Nasdaq Global Select Market |
| Ticker | ANT |
| Lead underwriters | Goldman Sachs and Morgan Stanley |
| Price and size | Not yet set |
| Earlier step | Confidential filing announced on 1 June 2026 |
That earlier step matters for timing. Anthropic told the market it had filed confidentially on 1 June (Reuters, 2026). The public filing came almost five months later.
Citation capsule
On 26 October 2026, Anthropic publicly filed a Form S-1 with the US Securities and Exchange Commission to list its Class A common stock on the Nasdaq Global Select Market under the ticker ANT. The offering size and price range are not yet set, and the registration statement had not become effective, so no shares could be sold. (Anthropic, 2026)
Is Anthropic Making a Profit?
Not on the official numbers, though most of the headline loss is a paper revaluation.
The S-1 reports a net loss of $80.7 billion for the six months to June 2026. Two items in it are changes in what financial instruments are worth on paper. One is $72.4 billion from convertible notes, derivatives and SAFEs (simple agreements for future equity, a type of early investor deal). The other is $8.7 billion from other market-linked instruments. Together they come to $81.1 billion, which is more than the whole loss. These are valuation changes, not cash leaving the company.
Anthropic’s adjusted figures strip those items out, along with share-based pay and a few others. On that basis it reports adjusted net income of $2.7 billion for the half-year. Its operating loss was $693 million, against $2.1 billion a year earlier. “Adjusted” is the company’s own measure, not an accounting standard, so read it as Anthropic’s view of itself.
Why does a customer care? Because a supplier that is close to break-even on operations behaves differently from one burning cash. It is less likely to need a sudden price rise to survive.
Citation capsule
Anthropic reported revenue of $16.3 billion and a net loss of $80.7 billion for the six months ended 30 June 2026. The period included $72.4 billion of changes in the fair value of convertible notes, derivative liabilities and SAFEs, and $8.7 billion from other mark-to-market instruments. Its adjusted net income, a non-GAAP measure that excludes these and other items, was $2.7 billion. (Anthropic S-1, 2026)
Does the IPO Change My Risk as a Claude Customer?
No. The IPO adds no new risk, but it forces Anthropic to describe its risks in plain legal language you can read.
Some of those risks matter more to a customer than to a shareholder. Each comes with a question for your own team.
Risk 1: Can a Few Big Customers Sink the Company?
Not on the filing. No single customer was more than 10% of revenue in the nine months to 30 September 2026, and the top five were about 25% (Anthropic S-1, 2026).
The trend runs the right way. About 79% of third-quarter revenue came from outside the top five customers, up from about 64% in the first quarter. The filing also warns that many top customers have no long-term contract and can cut spending at any time.
So concentration is a smaller worry than you might expect. If your usage is small, do you have a plan for the day a larger account takes the capacity you rely on?
Risk 2: What Happens If Compute Runs Short?
Anthropic’s biggest cost is computing power it has already promised to pay for, so its prices and limits follow that bill.
The filing says Anthropic relies on Amazon, Google and Microsoft for compute. Its long-term deals include “take-or-pay provisions” that “obligate us to pay for resources regardless of actual utilization.” Substantially all of its committed spend is compute and infrastructure. The dollar total is left blank in this preliminary version, so I will not quote one.
There is a good sign for customers. Anthropic says Claude Opus 5.5 cut customer costs per task by about 40% compared with Claude Opus 5. That is the company’s claim, and I have not tested it. Which of your workflows would break if rate limits tightened or prices moved?
Risk 3: Can a Government Decision Switch a Model Off?
Yes, and it has already happened once.
The S-1 says that in June 2026 the US Department of Commerce “imposed worldwide export restrictions” on the Fable 5 and Mythos 5 models. Anthropic “disabled” them “for all customers until those restrictions were lifted.” They came back after the company updated its safeguards.
The filing describes a second dispute. In March 2026 the Department of War sent letters asserting that Anthropic’s models present a supply-chain risk. It also warns that limits on selling to customers outside the United States “may cause our revenue from such unsupported jurisdictions to decline.”
If you run a business in Asia-Pacific, this is the risk that matters most. The model you use can change overnight for reasons that have nothing to do with how good it is. If your main model vanished on Monday, who on your team would know what to do by Tuesday?
Citation capsule
In its S-1, Anthropic states that no single customer exceeded 10% of revenue and its top five customers were about 25% for the nine months to 30 September 2026. It also discloses that in June 2026 the US Department of Commerce imposed worldwide export restrictions on its Fable 5 and Mythos 5 models, which it disabled for all customers until the restrictions were lifted. (Anthropic S-1, 2026)
I read the filings and launch notes so you do not have to, and write up only what changes a buying decision. Join the newsletter and the next one lands in your inbox first.
Does Anthropic’s Governance Structure Matter to a Business Buyer?
A little. It tells you who can overrule the revenue team, which is a fair question to ask of any AI supplier.
Anthropic is a Delaware public benefit corporation. That means the board must balance shareholders’ money interests, the interests of people affected by the company, and a stated public purpose. It has also set up a Long-Term Benefit Trust. Through a special class of stock, the Trust can elect and remove directors who together will be a majority of the board. The filing calls Anthropic a “controlled company” under Nasdaq rules and says its multi-class structure may keep some institutional investors out.
For a customer, the useful point is narrow. A board with a safety purpose and a trust behind it may refuse a commercial opportunity that a normal company would take. That can protect you, and it can also mean a feature you want never ships. Neither is a reason to leave. Both are worth knowing before you build a process on top.
Is OpenAI Any Safer as a Vendor?
Nothing public says so yet. You can read Anthropic’s risks because it filed, and you cannot yet read OpenAI’s.
OpenAI is reported to have submitted a confidential draft registration in early June 2026 (MediaNama, 2026). As of 28 October I found no public OpenAI S-1, so check EDGAR for the current position. Anthropic’s own filing names OpenAI Group PBC and Google as its primary competitors.
The three questions apply to every model supplier. How much of its revenue rides on a few accounts? How much compute has it promised to buy? What happens to your access if a government steps in? Ask them of OpenAI and Google too. In my survey of Asia-Pacific startups, most firms already used two or all three tools, which is the best protection a small company has.
What Should a Business Do This Month?
Make your dependence on one model visible, then make it reversible. None of this needs a lawyer or a budget.
- List every workflow that calls Claude, and mark the three that would hurt most if it stopped.
- For those three, name a second model and run the workflow on it once. A swap you have never tested is not a plan.
- Keep prompts, skills and workflow files in your own folders, not only inside one vendor’s app. I explain how I set this up in Claude Code for marketing.
- Read the “Risk Factors” section of the S-1 once. Start at page 26. It is long, but the sections on customers, compute and government are the ones above.
- Put a human approval step in front of any agent that can spend money or send email. My longer guide is AI agents for business in 2026.
Can I Buy Anthropic Shares?
Not yet. Shares can be bought only after the listing, through an ordinary brokerage account.
Until then, prices you see on private-market sites are not IPO prices. Whether you should buy is a question for a licensed financial adviser, not for a post about vendor risk. This article makes no view on the valuation.
When OpenAI's S-1 goes public, I will read it the same way and send the customer-risk summary to subscribers before it goes on the site.
Join the newsletter
Frequently Asked Questions
What is the Anthropic ticker symbol?
ANT. Anthropic has applied to list its Class A common stock on the Nasdaq Global Select Market under that symbol. The listing has not happened yet, so the ticker is not live and no shares can be bought.
When will Anthropic go public?
Press reports expect a listing in November 2026, but the company has not set a date and reports differ on the week. Anthropic’s S-1 says the registration statement had not yet become effective when it was filed on 26 October 2026, so nothing can be sold until it does.
Is Anthropic profitable?
Not on the statutory accounts. The S-1 shows an $80.7 billion net loss for the six months to June 2026, alongside $81.1 billion of paper revaluation charges on notes and similar instruments. On Anthropic’s own adjusted measure, which excludes those items and others, it reports $2.7 billion of adjusted net income for the same period.
Will the IPO change Claude's pricing?
The filing does not announce any price change, and I would not predict one. What it does show is that Anthropic has large fixed compute commitments, so its prices and usage limits depend on how well it manages that bill. It also says Claude Opus 5.5 cut customer costs per task by about 40% against Claude Opus 5.
Can the US government stop me using Claude?
Government action has already led to two Claude models being switched off once. In June 2026 the US Department of Commerce imposed worldwide export restrictions on Fable 5 and Mythos 5, and Anthropic disabled them for all customers until the restrictions were lifted. Keep a tested second model for any workflow you cannot afford to lose.
Should I move from Claude to ChatGPT or Gemini because of the IPO?
No. The filing makes Anthropic’s risks readable, not larger, and the same three questions apply to every supplier. A better move is to keep a second model tested for your most important workflows, so a change by any one vendor, or any one government, does not stop your work.
More on AI for business: AI for Business | Related: APAC AI Adoption 2026 report | ChatGPT vs Claude vs Gemini in Asia-Pacific | OpenClaw vs Claude Managed Agents