Anthropic made its U.S. IPO filing public on October 26, 2026. The maker of Claude plans to list on Nasdaq under the ticker ANT. The documents reveal just how fast the AI company is growing, the enormous financial commitments it is taking on and how much control ordinary shareholders will have.
In its official announcement of the filing, Anthropic confirmed it plans to offer Class A common stock. The price per share and number of shares have yet to be set. A definitive first day of trading has not been announced either.
The filing offers a much fuller picture of the company. Anthropic generated $16.3 billion in revenue in the first half of 2026 and posted an operating profit in the second quarter. At the same time, it has signed contracts worth hundreds of billions of dollars for computing capacity and infrastructure over the coming years.
What do we know about Anthropic’s IPO?
Anthropic has filed a Form S-1 with the Securities and Exchange Commission, the U.S. securities regulator. A company uses this registration statement to disclose information about its business, financial results, shareholders and risks before offering shares to the public.
IPO stands for initial public offering: the first public sale of a company’s shares. Publishing an S-1 is a major step in that process. The IPO itself has not yet taken place.
| Item | Status as of October 27, 2026 |
| Company | Anthropic, PBC, maker of Claude |
| Planned exchange | Nasdaq Global Select Market |
| Planned ticker | ANT |
| Shares to be offered | Class A common stock |
| Offer price and number of shares | Not yet determined |
| Final IPO valuation | Not yet known |
| First day of trading | Not officially announced |
Anthropic had already submitted a confidential draft on June 1. AI Wereld reported on the first formal step toward an IPO at the time. The public filing now makes the documents accessible to a much wider audience.
According to The Wall Street Journal, Anthropic is aiming for the week of November 16. Citing people familiar with the matter, the newspaper reported a possible raise of $100 billion or more and a valuation of $1.5 trillion to $2 trillion. Those figures and the timeline are not yet confirmed terms of the offering.
Goldman Sachs and Morgan Stanley are leading the offering as underwriters. Anthropic says the registration statement has not yet become effective. The announcement therefore does not mean the company has completed its Nasdaq listing.
How fast is Anthropic really growing?
Reported revenue from January through June 2026 was about 14 times higher than a year earlier. The figures in the public SEC prospectus put that growth into perspective.
| Financial metric | First half of 2025 | First half of 2026 |
| Revenue | $1.163 billion | $16.286 billion |
| Computing and infrastructure costs | $2.334 billion | $8.945 billion |
| Total operating expenses | $3.298 billion | $16.979 billion |
| Operating result | $2.135 billion loss | $693 million loss |
| Net result | $12.969 billion loss | $80.748 billion loss |
The half-year figures are unaudited interim results. The 2023, 2024 and 2025 annual figures in the prospectus have been audited by an independent accounting firm.
The second quarter changed the financial picture significantly. Anthropic generated about $11.6 billion in revenue and $974 million in operating profit that quarter, as The Next Web reports, citing the prospectus. The company still posted an operating loss for the half-year as a whole because the first quarter was in the red.
Operating profit means revenue exceeded operating expenses during that period. Interest, certain financial revaluations and taxes come afterward. That is how a company can make an operating profit yet still report a large net loss at the bottom of its income statement.
The results offer an early sign that Anthropic may be able to deliver its services profitably at scale. But one profitable quarter does not prove that future models, customers and infrastructure will maintain the same balance between revenue and costs.
Revenue run rate is not the same as revenue earned
Earlier reports about Anthropic often cite its annualized revenue run rate. That figure takes the company’s recent revenue pace and projects it over a year. For example, $1 billion in revenue in a single month would imply an annual run rate of $12 billion. It does not mean the company has already earned that amount.
When Anthropic announced its funding round in May, it said its revenue run rate had topped $47 billion a year. AI Wereld later reported a $65 billion run rate cited by Reuters. These snapshots cannot be directly compared with revenue from a completed half-year or a full-year financial statement.
The reported $16.8 billion in third-quarter revenue also needs a separate label. Reuters reported that figure after a closed-door investor meeting. The relevant revenue fields in the newly published S-1 have not yet been completed. That gives the figure a different status from the revenue in the published half-year table.
Why is Anthropic reporting an $80.7 billion loss?
Most of the net loss is tied to Anthropic’s financing. About $72.4 billion stems from the revaluation of convertible debt, derivatives and other financing arrangements.
A convertible loan allows a lender, under certain conditions, to turn the debt into shares. If the company’s value rises sharply, the accounting value of such a financing instrument can rise too. Anthropic records that change as an expense on its income statement.
There is no equivalent cash payment at that point. So the $80.7 billion net loss does not mean Anthropic spent that amount in cash on chips, salaries and data centers in six months. AI Wereld previously explained the distinction in its coverage of the huge loss that was largely on paper.
But it would also be misleading to subtract only the $72.4 billion from the net loss and present the remainder as a loss from operations. Other financial items also affect the net result. The operating loss for the half-year is reported separately: $693 million.
Accounting losses are not automatically meaningless, either. The underlying agreements may entitle investors to shares or future payments. For shareholders, both cash flow and the question of how much of the company ultimately goes to existing and new investors matter.
Anthropic also publishes adjusted profit measures that exclude certain items. These figures show how the company performs without those items, but the nature and scale of the excluded costs still matter. Stock-based compensation, for example, does not require an equivalent cash payment upfront, but it can dilute other shareholders’ ownership.
Why does Anthropic need to raise more money?
As of June 30, Anthropic had about $112.4 billion in cash and cash equivalents, plus $19.3 billion in short-term investments. That comes to roughly $131.7 billion, according to MarketWatch’s reading of the prospectus. This is the company’s position on that date, not its current bank balance in October.
The company reported positive operating cash flow of $3.929 billion for the first half. The timing of customer payments and payments to suppliers also played a role, according to the financial disclosures.
Cash flow tracks money actually coming in and going out. A customer may pay in advance for services the company will deliver later; a supplier may not be paid until a later date. Positive operating cash flow, on its own, therefore does not prove that day-to-day operations can fund all future investment.
Anthropic also has to secure capacity before the revenue tied to it comes in. Chips, data center space, power and long-term leases all require planning and financial commitments. Raising more money could give the company room to take on those obligations and absorb setbacks.
The proposed IPO includes both newly issued shares and shares sold by existing shareholders. Only proceeds from the new shares go to Anthropic. So a huge headline figure for the IPO does not automatically mean the company will receive the same amount for research and infrastructure.
How does Anthropic make money from Claude?
Anthropic was founded in 2021. Co-founders Dario Amodei and Daniela Amodei lead the company as CEO and president, respectively.
Anthropic sells access to AI models and applications. Businesses can integrate Claude into their own software through an API. Users can also work directly with Claude and products such as Claude Code, which helps with software development.
Customers can pay in different ways. A subscription provides access to certain features and usage limits. With usage-based pricing, the bill depends on the models selected and the amount of information processed. A company using Claude in thousands of daily workflows can therefore spend far more than an individual chatbot user.
Claude Code is an important driver of that business growth. In February, Anthropic said Claude Code had reached an annual revenue run rate of more than $2.5 billion. Business use accounted for more than half of that product’s revenue at the time. This is a historical snapshot, not Claude Code’s current annual revenue.
Anthropic’s customer base now includes major buyers. According to the AP’s review of the IPO documents, as of September 30, 2026, Anthropic had about 250 customers that each generated at least $10 million in revenue over the previous 12 months. At the same time, its five biggest customers accounted for about a quarter of revenue in the first nine months of 2026.
That concentration can accelerate growth when major customers expand their use. But it also makes revenue more vulnerable to decisions made by a small group of buyers. If a major customer develops its own model, uses less AI or switches to a competitor, the impact could show up in Anthropic’s revenue.
Where does the reported $2 trillion valuation come from?
The reported IPO valuations build on the rapid increase in Anthropic’s value across earlier funding rounds. The company announced the following rounds:
| Date | Funding round | Amount raised | Post-money valuation |
| March 3, 2025 | Series E | $3.5 billion | $61.5 billion |
| September 2, 2025 | Series F | $13 billion | $183 billion |
| February 12, 2026 | Series G | $30 billion | $380 billion |
| May 28, 2026 | Series H | $65 billion | $965 billion |
A valuation is the value assigned to a company in a transaction. The amount raised is the money investors put in. So Anthropic was not valued at $65 billion in its May round: its agreed post-money valuation was $965 billion.
The Series H round also included $15 billion in previously committed investments from major cloud companies. Those amounts should not be counted again as separate new funding rounds.
A possible IPO valuation of $1.5 trillion to $2 trillion would mark another major leap. In English, a trillion is 1,000 billion. The reported range therefore equals $1,500 billion to $2,000 billion.
Such a price assumes Anthropic can keep creating substantial economic value for years to come, even after paying the cost of that growth.
A high company valuation also tells us nothing about the price of a single share. That requires the total number of shares after the IPO, including any relevant conversions of existing financing. Until those figures and the offer price are known, it is impossible to make a precise assessment per share.
The biggest financial question lies in its compute contracts
Anthropic has secured capacity on a massive scale for years to come. In April, for example, the company announced plans to spend more than $100 billion on AWS technology over ten years. The prospectus now gives more detailed figures for several agreements.
| Partner | Amount in contract description | Period or key condition |
| Amazon | At least $110 billion | May 2026 to April 2036 |
| At least about $114.5 billion | April 2026 to July 2033 | |
| Microsoft | At least $44.9 billion | November 2026 to May 2033 |
| Broadcom | About $161.2 billion, excluding interest | Equipment leases with five-year terms |
| SpaceXAI | Up to $120.5 billion | Contracts running through 2029, with termination options after initial fixed periods |
These figures cover different kinds of agreements. The cloud contracts cited generally include minimum purchase commitments, even if Anthropic uses less capacity. The SpaceXAI agreements, by contrast, include termination options after the initial fixed periods. Adding all the figures together and calling the result current debt would therefore give a false impression.
Capacity also comes online at different times, depending on the project. Anthropic expects new capacity through Google and Broadcom to become available from 2027. The company uses several kinds of chips, including Google TPUs, AWS Trainium chips and Nvidia GPUs.
AI companies face two major types of computing workload. Training builds and improves a model. Inference is the work the model does to respond to user requests. A successful service therefore needs to keep buying or providing computing power even after the model has been developed.
Long-term contracts give Anthropic more certainty about access to capacity. But they also carry financial risk. If customers spend less than expected, fixed purchase commitments may still have to be met. If new chips become much cheaper, previously secured capacity could look relatively expensive.
On the other hand, a shortage of capacity can limit growth. In May, Anthropic linked additional computing power through SpaceX to higher Claude usage limits. That shows how directly infrastructure affects the product customers receive.
When assessing these contracts, the annual payment schedule, minimum purchases, start conditions and termination rights matter more than one eye-catching total.
Amazon, Google and Microsoft wear several hats
For Anthropic, major tech companies are investors, suppliers and sales channels all at once. Some also offer competing AI products. The combination makes these commercial relationships powerful—but complicated.
Amazon announced an additional $5 billion investment in April, with up to $20 billion in potential additional investments, subject to certain conditions. At the same time, Anthropic agreed to purchase large amounts of AWS capacity over the long term.
The partnership with Microsoft and Nvidia also brought together access to Azure, technical collaboration and planned investments. The Azure commitment announced in November 2025 was $30 billion; the current contract description in the prospectus now cites a higher minimum.
The economic link is clear: an investment helps Anthropic grow, and the company then spends some of that money with infrastructure providers. That can benefit both sides. It does not, on its own, prove that revenue is artificial or that customers do not want the products.
But it does make the source of end-user demand crucial. For the business model to last, customers must see enough value to cover the costs across the entire chain. If financing remains available while demand falls short, the gap between promised growth and actual returns could widen.
Who will hold the power after the IPO?
New shareholders will gain economic rights, but Anthropic is not simply handing over control. The company has a distinctive structure designed to protect the founders’ role and its public-benefit mission.
Under the IPO charter, the founders’ corporation will receive a special Class F share. That share carries 50.1% of the general voting power. The founders also hold common shares.
That 50.1% does not describe their economic ownership. The special share itself carries no ordinary claim on profits or dividends. Its purpose is to secure voting control. A publicly traded Class A share carries one vote.
The Long-Term Benefit Trust also has a distinct role. Under the shareholders’ agreement, the trust generally has the right to appoint a majority of the board. One of those board seats is subject to an additional approval process. The trust does not run the company’s day-to-day operations.
Anthropic is a public benefit corporation, a U.S. corporate structure that combines a for-profit business with a defined public-benefit purpose. Anthropic says it aims to develop advanced AI responsibly for the long-term benefit of humanity. In April, the company said the directors chosen by the trust already made up a majority of the board.
For shareholders, that protection cuts both ways. It may give the company room to defend safety decisions and long-term investments, even when they come at a short-term cost. But ordinary shareholders will have less power to change course if they disagree with the board and the founders.
A separate fund for public-benefit work
On October 18, Anthropic also announced the Anthropic Public Benefit Fund. Subject to shareholder approval, the company plans to allocate 5% of its shares to the fund. The board may add more shares if certain valuation milestones are reached.
The fund is separate from the Long-Term Benefit Trust. The trust safeguards the governance structure; the fund is intended to support public-benefit work. The fund’s detailed structure and valuation milestones have not yet been fully disclosed.
The founders’ pledge to dedicate 80% of their own shares to philanthropic causes also concerns their personal holdings. It does not mean that 80% of Anthropic has already been transferred to charities.
Which risks go beyond the balance sheet?
Anthropic has to keep defending its technological edge. A model that looks compelling today may face competition tomorrow from a cheaper or better alternative. Customers may also use several models side by side. Rapid adoption does not automatically mean lasting dependence on one provider.
Copyright remains a concrete financial risk. On July 20, 2026, a U.S. judge gave final approval to a $1.5 billion settlement with book rights holders. The settlement covers defined claims relating to the earlier acquisition and copying of works. It does not resolve every possible claim involving future activities or AI-generated output.
AI safety also affects the business. In its September 2026 threat report, Anthropic describes attempts to use Claude for cyberattacks, surveillance and fraud, among other activities. The company says it disrupted the activity described. This is Anthropic’s own assessment, not an independent measure of the safety of every application.
Misuse, serious errors or a loss of control could lead to lawsuits, product restrictions and loss of trust. A risk disclosure in an IPO filing does not mean a described disaster is certain to happen. It does show which potential consequences the company considers relevant to shareholders.
Disputes with governments can also affect the business. In October, Anthropic sought a rehearing of a ruling over restrictions on U.S. defense contracts. The FTC also demanded information on consumer and data protection, among other topics, Reuters reported, citing the IPO documents. An information request is not a finding of wrongdoing.
Can retail investors buy Anthropic shares yet?
Anthropic has not completed its public listing under the ticker ANT. The IPO filing therefore does not mean anyone can buy the stock on Nasdaq yet.
Subscribing for shares in an IPO is also different from buying them once trading begins. A subscription is a request for an allocation of shares at the offer price. After trading starts, the market sets a price that can move up or down.
The prospectus provides for a possible offer to retail investors in certain countries in the European Economic Area through a separate European prospectus. The countries involved have not yet been named. Participation by investors in the Netherlands has therefore not been confirmed.
Key details are still missing: the offer price, the final number of shares, the split between new and existing shares, and further updated financial information. The timing of when existing shareholders are allowed to sell more shares could also affect the supply of stock on the market later.
What does the IPO still need to prove?
The next documents should show the terms on which new shareholders can invest. After that, attention turns to execution: will the capacity Anthropic has committed to generate enough paid work, and will margins hold up as new investments start to weigh on the business?
For Claude users, that comes down to practical factors such as availability, usage limits, price and product quality. Business customers will also care about the reliability of their provider. For shareholders, the ultimate question is how much revenue remains after all obligations are paid.
A successful IPO could give Anthropic a major financial boost. The harder test comes afterward: turning years of committed computing capacity into services that customers will keep paying for, year after year.
Popular news
Akamai launches Agent Sandbox to limit what AI code can do
More FPS with AI and DLSS Doesn’t Automatically Mean a More Responsive Game
OpenAI Launches GPT 6.1 Terra for Faster AI Agents at Half the Price
Anthropic's IPO Filing Reveals $16.3 Billion in Revenue and the Cost of Growth
Microsoft Restructures Workforce with Voluntary Buyouts Amid AI Expansion
Microsoft Strengthens European Digital Sovereignty With New Cloud Controls
Latest comments
Loading