Anthropic prospectus: $65 billion ARR, $80 billion net loss, expected IPO valuation lowered to $1.5 trillion

[[**Anthropic's prospectus shows ARR of $65 billion, adjusted profitability in the first half but an operating margin of only 10%, carrying roughly $850 billion in compute bills, with expected IPO valuation cut to $1.5 trillion.**

**Key points**

**1. Revenue growing fast but dependent on new models**

2026H1 revenue was $16.286 billion, up about 13x year-over-year, with API accounting for 80%; models released in the past six months contributed over 80% of revenue in the most recent three months, and a slowdown in updates would hit growth.

**2. Huge loss comes from valuation remeasurement; adjusted profit remains thin**

2026H1 net loss was $80.748 billion, mainly a $72.388 billion fair value change loss; adjusted operating margin was only 10%, with gross margin at 85%.

**3. $850 billion compute bill locks in cost inflexibility**

Long-term agreements with Broadcom, SpaceX AI, Amazon, Google and others total about $850 billion, mostly purchased 12 to 18 months in advance; forecast errors will lead to idle capacity or missed opportunities.

**4. Special governance and no-advertising restriction limit profit paths**

The company is a public benefit corporation, with governance rights held by a long-term benefit trust and founder shares; it has committed that Claude will not carry ads, so profitability depends on API and subscriptions.

**5. IPO valuation expectations pulled back, providing an industry pricing anchor**

The most recent financing had a post-money valuation of $965 billion, about 20x P/ARR; the IPO valuation expectation discussed by investors fell from about $2 trillion to $1.5 trillion, while domestic peers are at about 10x and 7.8x. 2026-10-27 10:08 $65 billion ARR, $80 billion net loss, the strongest AI company lifts the veil Dingjiao One quick look

This article comes from the WeChat public account: Dingjiao One, author: Dingjiao One team, editor: Wei Jia

Burning $200 million a day, who is Anthropic working for?

Dingjiao One (dingjiaoone) original

Anthropic may not be the largest IPO in history, but it is certainly one of the most anticipated. On the evening of October 26 Beijing time, people waiting for Anthropic to publish its S-1 prospectus directly crashed the U.S. Securities and Exchange Commission (SEC) website.

The company was founded only five years ago. In 2021, seven former OpenAI members left to start the company; five years later, Anthropic has overtaken its former employer in both valuation and revenue. In May this year, it surpassed OpenAI with a post-money valuation of $965 billion, becoming the world's most valuable AI startup. At the end of July, its ARR reached $65 billion, compared with about $40 billion for OpenAI in the same period.

The prospectus shows that in the first half of 2026, Anthropic's revenue grew about 13x year-over-year to $16.286 billion, of which revenue billed by API usage accounted for 80% and subscription revenue 20%. In the same period, the book net loss exceeded $80 billion, but this was mainly affected by fair value changes in financing instruments; on an adjusted basis it was already profitable, with an operating margin of 10% and a gross margin of 85%.

Shortly before this prospectus was made public, Reuters reported that because the company's quarter-over-quarter revenue growth had slowed, some investors attending Anthropic's investor day discussed lowering the expected IPO valuation from the previous $2 trillion to about $1.5 trillion. This figure is only an expectation in investor discussions, and the final offering valuation has not yet been determined. If issued at a $1.5 trillion valuation, it would be below SpaceX's $1.77 trillion valuation at its IPO in June this year.

Investors' concerns are not unfounded. Even Anthropic, which is running at the front, is not yet enough to prove that large models are a good business.

High compute costs, a lead that could shrink at any time, and long-term returns that have not yet been validated may all change the profit curve of this business. The answer given by the prospectus is also far more complex than an adjusted-profit report card.

01. The most profitable large-model company has an operating margin of only 10%

Several brokerage analysts who have long followed U.S. tech companies told "Dingjiao One" that they take a cautious view of Anthropic's financial data. This adjusted-profit prospectus not only fails to strengthen market confidence, but instead exposes problems in the business model of frontier model companies.

The first problem is reflected in the revenue structure and growth method.

The prospectus shows that from 2023 to 2025, Anthropic's revenue increased from $41 million to $4.585 billion, a two-year compound growth rate of about 957%. Revenue in 2024 and 2025 was about 9.4x and 11.9x that of the previous year respectively, equivalent to revenue scale roughly doubling 10 times a year on average.

In the first half of 2026, Anthropic's revenue scale further increased to $16.286 billion, up about 1,300% from $1.163 billion in the first half of 2025. At least judging from the data in the prospectus through the end of Q2 2026, the revenue growth slowdown investors care about has not yet occurred.

The change in ARR (annualized revenue) behind the revenue figures better reflects the frightening speed of Anthropic's growth. At the start of 2025, Anthropic's ARR was only $1 billion, and by the end of the year it was roughly $9 billion. In April this year, Anthropic officially disclosed that the company's ARR had surpassed $30 billion, exceeding the $25 billion OpenAI had disclosed to the public two months earlier.

Enterprise customers are the main factor driving the company's growth.

The prospectus shows that as of September 30, 2026, Anthropic's customer base included more than 11,900 customers that contributed $100,000 or more in revenue over the past 12 months (TTM); more than 1,900 customers that contributed $1 million or more; and about 250 customers that contributed $10 million or more.

In terms of revenue types, Anthropic's revenue is mainly divided into two categories: API usage-based billing and subscriptions. Usage-based revenue increased from $288 million in 2024 to $3.8 billion in 2025, with its share rising from 75% to 83%, and in the first half of 2026 it further reached $13.1 billion, with its share slipping slightly to 80%. Over the same periods, subscription revenue accounted for 25%, 17% and 20% respectively. Usage-based billing occupies the main position, and revenue is closely tied to how frequently and at what scale customers continue to call the models.

The prospectus also discloses a detail that is easy to overlook. Model versions released in the six months ended September 30, 2026, contributed more than 80% of revenue in the three months ended on that date.

A USD fund investor told "Dingjiao One" that this is fairly common in the large-model field, with revenue growth largely driven by new models. A company must continuously update versions and keep leading in order to sustain revenue growth. Once it goes a long time without updating, or a new model falls slightly behind, revenue could take a significant hit.

Another hidden risk in the prospectus is reflected in the profit margin.

From 2023 to 2025, Anthropic's net losses were $1.531 billion, $8.309 billion and $41.965 billion respectively. In the first half of 2026, the net loss widened to $80.748 billion, about 6.2 times the $12.969 billion in the same period the previous year.

This huge loss needs to be broken down. In the first half of 2026, the fair value change loss on convertible notes, derivative liabilities and SAFEs reached $72.388 billion. A rise in the company's valuation pushes up the fair value of related liabilities, creating a non-cash loss. The principle is similar to the fair value losses on preferred shares commonly seen at Hong Kong IPO companies, but the instrument structure and accounting treatment are different. The related remeasurement will stop only after the instruments are converted, repaid or settled, and this is not a loss generated at the operating level.

At the operating level, break-even has already been achieved. In the first half of 2026, Anthropic's GAAP operating loss was $693 million, adjusted operating profit was $1.618 billion, the adjusted operating margin was 10%, and adjusted net profit was $2.671 billion.

This is basically consistent with data previously reported by foreign media. In August this year, foreign media reported that Anthropic achieved positive adjusted operating profit for the first time in the second quarter. At its investor day in October, the company also disclosed third-quarter revenue of $16.8 billion, up about 45% quarter on quarter, achieving adjusted profitability and positive cash flow for the second consecutive quarter.

As can be seen, Anthropic's profit margin is not ample, and profitability depends mainly on rapid revenue expansion, as well as on computing efficiency and expense control. Scale growth in turn bets on frequent model updates and sustained leadership. Once the pace of model capability improvement slows, revenue, profit and subsequent R&D investment may produce a chain reaction.

Of course, large-model companies do not have only one business model, selling APIs. Embedding advertising is a high-margin business model, and OpenAI is doing it. But Anthropic promised in its prospectus to keep Claude free of advertising at all times, explaining that advertising could bring considerable revenue, but the company wants Claude's only incentive to be providing helpful answers, avoiding pressure from revenue targets to pursue user engagement.

This choice preserves the product positioning, and it also means that at this stage Anthropic still needs to rely on model calls and subscription revenue to prove its profitability.

02. Carrying a computing bill of more than $800 billion, locking in the next decade

At this stage, AI large models are still not an easy business, and Anthropic is well aware of this. That is why the special feature of this prospectus is that it uses a large amount of space at the beginning to discuss AI safety and the company's unique equity setup to avoid the risk of AI going out of control, trying to explain how it maintains long-term goals beyond commercial returns.

Because Anthropic was registered as a public benefit corporation (PBC) when it was founded, it is legally required to consider simultaneously the interests of shareholders, the groups affected by the company's actions, and the public benefit goal of responsibly developing and maintaining advanced AI. This arrangement brings safety goals into the company's articles of association, but it cannot automatically eliminate commercial pressure, nor can it guarantee that the company will always prioritize safety.

In terms of governance structure, the Long-Term Benefit Trust (LTBT), through Class T common stock, generally has the right to elect and remove Class T directors holding a majority of the authorized board seats, and Class T carries no economic rights. The Founder LLC established by the seven co-founders will receive one share of Class F common stock before the listing, corresponding to 50.1% of the voting power on general shareholder matters, likewise carrying no economic rights.

The public benefit objective and founder control are both written into the post-listing structure, and public shareholders need to accept this special allocation of power. This structure answers the question of whom the company wants to safeguard its long-term mission, but whether the mission can be delivered depends on whether the company has enough money to keep investing; model R&D, talent retention, and compute procurement all require long-term funding.

The size of the fundraising in this listing and the specific amounts for each area of investment have not yet been disclosed. The prospectus lists the general uses as working capital, operating expenses, and capital expenditures; capital expenditures cover compute and infrastructure investment, and the funds will also be used to recruit and retain highly skilled talent. The company may also acquire or invest in related businesses, products, services, and technologies, but there are currently no material acquisition or investment agreements or commitments.

On talent, as of September 30, 2026, Anthropic had approximately 5,300 employees, with revenue over the past 12 months exceeding $6.6 million per employee, clearly above the average of the largest listed companies in the United States.

On compute, the prospectus discloses multiple long-term agreements. The company's five-year equipment lease rent related to Broadcom is approximately $161.2 billion, excluding interest; rental payments to SpaceX AI for NVIDIA GPU compute amount to as much as $120.5 billion; and service commitments to Nscale total up to approximately $44.6 billion.

Beyond the above contracts, Anthropic has also committed to paying Amazon at least $110 billion between 2026 and 2036, and Google approximately $114.5 billion between 2026 and 2033, with an additional estimated $125 billion to $150 billion in TPU compute spending. Long-term arrangements related to Microsoft and Lambda also amount to $44.9 billion and approximately $41.3 billion, respectively. The different agreements are inconsistent on minimum payments, maximum spending, delivery conditions, and cancellation terms, and the compute-related bills disclosed in the prospectus total approximately $850 billion.

These contracts change the company's cost flexibility. The prospectus states that most compute is procured 12 to 18 months in advance, costs are recognized once compute becomes available, and near-term spending is relatively fixed. Demand forecasts that are too low may miss revenue and R&D progress, while forecasts that are too high may create idle capacity, and hardware and facilities locked in today may lose their cost advantage several years from now.

Anthropic's relationships with cloud providers are therefore more complex. Amazon, Google, and Microsoft are simultaneously shareholders and compute suppliers, and may also be customers, distribution channels, and backers of competitors. They provide Anthropic with capital and infrastructure, while Anthropic locks in future spending through long-term procurement commitments. A listing can replenish funds, but it cannot detach this business from the input network made up of talent, chips, cloud providers, and long-term contracts.

03. Pricing large models now has a reference point

Anthropic's prospectus is not only a performance showcase for a frontier model company; it also provides the large-model industry with a new valuation anchor.

Looking back at Anthropic's valuation changes, several milestones in 2026 are particularly key.

In February, Anthropic completed a $30 billion Series G financing at a post-money valuation of $380 billion. The company disclosed ARR of approximately $14 billion at the time, with Claude Code contributing $2.5 billion; in April, on the private secondary market platform Forge Global, Anthropic's implied valuation briefly reached $1 trillion, while OpenAI was at about $880 billion over the same period. In the same month, Anthropic disclosed that ARR had surpassed $30 billion.

Just one month later, Anthropic completed a $65 billion Series H financing at a post-money valuation of $965 billion, the most recent publicly disclosed valuation before the company's listing. According to Bloomberg, Anthropic disclosed to investors that by the end of July its ARR had risen to $65 billion.

Since the final IPO market capitalization cannot yet be determined, using the post-money valuation of the most recent financing and the corresponding annualized revenue as a reference, Anthropic's valuation-to-ARR ratio is approximately 20x.

By comparison, among the two already-listed model companies in the Chinese market, Zhipu disclosed ARR of approximately $1.8 billion in September, with year-end guidance of $3 billion; MiniMax's ARR exceeded $800 million in August, with a year-end target of $1 billion. Moonshot AI's ARR in August was reported to have exceeded $1 billion, with a year-end target of $2 billion.

Based on the closing prices on October 26 and year-end ARR expectations, the corresponding P/ARR for Zhipu and MiniMax is approximately 10x and 7.8x, respectively, far below Anthropic's 20x.

However, it is also important to note that the revenue types and statistical periods of ARR vary across companies. Whether API, subscriptions, on-premise deployment, customized services, and cloud revenue sharing are included differs among them. Large model companies generally have two types of ARR calibers: pure API revenue and MaaS revenue. Cloud vendors commonly use MaaS revenue; the difference is that MaaS includes API revenue as well as revenue from things such as model fine-tuning, on-premise deployment, and customization.

Zhipu's early disclosed ARR was pure API revenue, while the $1.8 billion disclosed in September was on a full-business caliber; Alibaba Cloud commonly uses MaaS revenue.

Another difference is the statistical period. Previously, both Zhipu and MiniMax specifically noted this in communications with investors: the former calculates on a monthly basis, while the latter calculates on a weekly basis. Anthropic's method is calculated monthly and counts API plus subscription revenue.

So where does the multiple gap mainly lie? Growth rate is not the key. The ARR of domestic companies is likewise growing rapidly; the gap is more reflected in scale and monetization efficiency per unit of computing power. Anthropic can already make money from computing power, while domestic companies are still working to make revenue cover computing power.

Looking further, computing power cost is the main expenditure of a model company. In accounting treatment, the computing power cost of early-stage model R&D is usually recorded as an expense and affects profit, while inference computing power is recorded as a cost and affects gross margin. Anthropic's gross margin in the first half of 2026 was as high as 85%, which is the unique advantage of high pricing and closed-source models.

Domestic large model companies mainly follow an open-source model and take a low-price route. With R&D expenses not differing much, the difference in gross margin directly widens the profit levels of the two.

But this yardstick itself is also shaking. Anthropic's expected IPO valuation has dropped from about $2 trillion to around $1.5 trillion; Zhipu's market value once reached HK$1.33 trillion intraday on June 22 and is now about HK$240 billion; MiniMax's stock price has fallen by more than 80% cumulatively from its March high. From the primary market to the secondary market, valuations of large model companies are all correcting.

Therefore, this is also a competition akin to the "Red Queen dilemma": "You have to run as fast as you can just to stay in place." Only by continuously updating the most cutting-edge models can this fragile business model be maintained. What this tests behind it is R&D capability, sustained cash-generation capability, and capital reserves.

This also explains why a company whose core narrative centers on safety and the public interest would actively seek to go public. Anthropic's offering results will become an important pricing case for frontier model companies and will also provide a reference for OpenAI, whose listing timeline points to 2027. Entering the listing process also means that for Anthropic, the focus is no longer limited to revenue growth brought by model leadership; it must also prove whether it can leave sufficient shareholder returns after long-term computing power investment.

*The cover image is from Unsplash. AI industry signal channel: Frontier Technology

572 articles have been published on Huxiu. This content is authorized by the author for publication; the views represent only the author and do not represent Huxiu's position. If you have objections or complaints about this manuscript, please contact tougao@huxiu.com. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]]]