Anthropic files S-1 to list on Nasdaq under the ticker ANT, with a valuation of up to $2 trillion

[[**Anthropic Files S-1 to List on Nasdaq Under "ANT" at a Valuation of Up to $2 Trillion, Potentially the Largest IPO in History; but Over $650 Billion in Compute Commitments and Slowing Growth Put the Trillion-Dollar Valuation to the Test.**

**Key Points**

**1. Sprinting Toward the Largest IPO in History**

Filed its S-1 on October 26, planning to list on Nasdaq under "ANT" at a valuation of $1.5 trillion to $2 trillion; if completed, it will surpass SpaceX to become the largest IPO in history, less than 5 years after the company was founded.

**2. Enterprise API Drives Revenue Surge**

Revenue in the first half of 2026 was $16.286 billion, up nearly 13-fold year over year, with a net revenue retention rate of 900%; about 80% of revenue comes from enterprise APIs, and there are nearly 12,000 customers spending more than $100,000 annually.

**3. Massive Compute Commitments and Slowing Growth Pressure the Valuation**

The company has signed long-term compute contracts worth more than $650 billion, about 80% of which are non-cancelable; compute spending in the first half of 2026 was $8.945 billion, already exceeding all of 2025. Q3 revenue was about $16.9 billion, with quarter-over-quarter growth slowing from 144% to about 47%, and some investors are discussing cutting the valuation from $2 trillion to about $1.5 trillion.

**4. Thin Profits at the Model Layer, With More Value Flowing to the Application Layer**

In 2026, hardware accounts for 93% of the AI industry's operating profit, while the model and application layers are close to zero, with improvement expected only by 2030; Dario acknowledges that abundant opportunities exist at the application layer, but believes that APIs and model iteration will continue to open up new application frontiers. 2026-10-29 08:05 5 years, 7 people, "overtaking" the former employer, sprinting toward the largest IPO in history Chinese Entrepreneurs Magazine Quick Read

This article is from the WeChat public account: Chinese Entrepreneurs Magazine, Editor: Ma Jiying, Author: Kong Yuexin

Can the capital and patience of the public market really support this "largest IPO in history"?

Dario Amodei is one of the most closely watched figures in the current AI industry.

On October 26, the Anthropic he founded officially filed its S-1, planning to list on Nasdaq under the ticker "ANT" at a valuation of $1.5 trillion to $2 trillion. By this scale, the listing will surpass the record set by SpaceX in June this year and become the largest IPO in history. From the company's founding to a market value of nearly $2 trillion, he took less than 5 years.

At the end of 2020, due to various disagreements with OpenAI's management, Dario left the company and founded Anthropic in early 2021. In March 2023, Anthropic released the first-generation Claude chatbot; in July of the same year, it launched Claude 2, which drew widespread market attention and acclaim.

But Dario did not focus on building consumer applications; instead, he mainly devoted himself to building AI for enterprise application scenarios. As demand for AI code generation exploded, Anthropic released the AI programming tool Claude Code in February 2025, and the company's revenue grew rapidly as a result. The prospectus discloses that in the first half of 2026, Anthropic's revenue was $16.286 billion, a year-on-year increase of nearly 13 times, and customers' net revenue retention rate reached 900%.

In just 5 years, Dario overtook his "former employer" OpenAI and will be the first to complete an IPO. But reality is not a wish-fulfillment story; behind the rapid growth, the company needs to spend heavily on R&D, computing power, infrastructure, and more. According to the prospectus, the company has already signed future computing power commitments exceeding $650 billion. Previously, due to pressure on the AI sector and the slowdown in the company's Q3 revenue quarter-on-quarter growth, some investors had discussed lowering Anthropic's valuation expectation from $2 trillion to about $1.5 trillion.

At a time when profit distribution across the AI industry chain is extremely unbalanced and giants are frantically taking on debt for infrastructure, can the capital and patience of the public market really support this "largest IPO in history"?

OpenAI's "Seven Defectors" Build Anthropic

The opening of The History of Venture Capital tells the story of the "Eight Defectors" collectively leaving Shockley Semiconductor Laboratory, ultimately giving birth to Fairchild Semiconductor and the entire Silicon Valley. More than half a century later, the same script is playing out in San Francisco. The birth of Anthropic also began with a Silicon Valley-style collective departure.

As a former vice president of research at OpenAI, Dario made enormous contributions to the R&D of GPT-2 and GPT-3, and was deeply involved in the early commercialization process.

But amid the rapid iteration of models, Dario began to question the direction of AGI. He repeatedly emphasized publicly that this technology must be advanced in an extremely cautious, direct, and honest manner, while OpenAI's increasingly aggressive pace of commercialization, as well as his disagreements with Sam Altman and other management over underlying values, led to a complete breakdown of mutual trust.

In December 2020, Dario decided to leave, and six of OpenAI's most core figures at the time also left with him, including his sister Daniela Amodei, then vice president of safety and policy operations, Tom Brown, first author of the GPT-3 paper, and physicist Jared Kaplan, who proposed Scaling Laws.

In February 2021, Dario and the others, determined to go it alone, founded Anthropic.

Dario was born in San Francisco in 1983. By his account, he was very close to his parents, so in 2006, influenced by his father's death from a rare disease, he switched from graduate studies in theoretical physics to the biophysics department. In 2011, Dario entered Stanford School of Medicine for postdoctoral study, researching neural circuits and tumor-related topics, but he soon realized that a single researcher could not achieve a breakthrough in this field within a limited time.

In November 2014, Dario joined Baidu's Silicon Valley AI Lab and, within the team of Andrew Ng (Baidu's former chief scientist), led the DeepSpeech 2 project, realizing a system approach that converts raw audio directly into text. This experience laid the groundwork for the intuition he later developed about Scaling Laws.

In 2015, he briefly joined Google Brain as a senior research scientist, but the pace and atmosphere of a big company made it hard for him to adapt, and he chose to leave in less than a year.

In July 2016, Dario joined OpenAI, which was initiated by Musk, Altman and others, and successively served as head of the AI safety research group and research director, eventually rising to vice president of research.

Dario's attention to AI safety began during his time at Google, where he also co-authored papers with colleagues exploring the potential bad behavior of AI. After the release of GPT-3, on the one hand he was excited by the technological progress, and on the other he became increasingly convinced that many things cannot be controlled simply by training models.

This philosophy began to create a split within OpenAI. Those who agreed with Dario quickly gathered around him and formed a close circle; but on the opposite side, management and some colleagues believed he was trying to use safety to "control" the company, and the ideological divide grew wider.

Dario's mentor Michael Berry once evaluated him this way: "He is one of the most talented students I have ever mentored, and also a rather proud person." This "pride" may be the reason he later broke completely with Altman and others. On the development (safety) of AGI, he refused to back down or compromise.

Dramatically, Musk, who had fallen out with Altman, opened fire directly on social platform X on February 13 this year, after Anthropic announced a $30 billion funding round, calling Claude "anti-human and evil." But less than three months later, on May 6, Anthropic officially announced that it would exclusively lease all the computing power of the Colossus 1 data center from SpaceX, including more than 220,000 Nvidia GPUs and 300 megawatts of computing power. Musk also changed his stance on X: "I talked with Anthropic's executive team, and I was impressed. No one triggered my evil detector."

On September 12 this year, Dario published a 3,800-word long article, "We Must Pace the Frontier," arguing that the industry must "slow down the pace of improving model capabilities" in exchange for one to two years to deploy safety safeguards.

Anthropic's prospectus also reflects Dario's emphasis on the safety narrative. Of the 261 pages of main text, the risk factors section occupies more than 80 pages, while the section describing business operations is only 48 pages.

Under the industry's model "arms race" and the performance pressure of going public, although Dario repeatedly emphasized slowing the pace of capability improvement, Anthropic has in fact officially released at least nine models this year, and has released five models in succession since September.

In fact, since 2025, Anthropic's model release pace has been continuously accelerating, updating about 10 versions of sub-models throughout the year, and its performance has also been soaring.

The prospectus shows that Anthropic's revenue has grown about 10-fold per year. From 2023 to 2025, the company's revenue was $41 million, $386 million, and $4.585 billion, respectively; revenue in the first half of 2026 was $16.286 billion.

Anthropic's choice of an enterprise-level (to B) business route has been validated by the market. The prospectus shows that in the first half of 2026, pay-as-you-go enterprise API calls contributed about 80% of revenue, and enterprise customers with annualized spending exceeding $100,000 surged from about 950 a year earlier to nearly 12,000.

Operating losses narrowed significantly. In the first half of 2026, Anthropic's operating loss was $693 million, compared with $2.135 billion in the same period the previous year; the operating loss margin fell from about 184% to about 4%. The company even achieved a quarterly operating profit in the second quarter of 2026, with operating profit of $974 million for the quarter and an operating margin of about 8%. After excluding items such as stock-based compensation and non-cash charitable donations, adjusted operating profit for the quarter was $2.235 billion, with an adjusted operating margin of 19%.

The technological progress and business performance since Anthropic's founding have kept it in the capital markets' spotlight. According to the prospectus, as of the end of June 2026, the company had raised a cumulative $141.1 billion. Tech giants such as Amazon, Google, Microsoft, and Nvidia, as well as several leading VCs, are among its shareholders.

However, based on public information and disclosures in the prospectus, since 2025 the investment logic of the big tech companies has fundamentally changed: capital investment is increasingly tied to compute procurement or chip supply, rather than being purely financial investment. The price Anthropic pays for accepting the funds is that, in the form of long-term procurement commitments, the money flows back to cloud providers and chipmakers.

The prospectus shows that the company's "compute and infrastructure" spending is expanding rapidly: $433 million in 2023, rising to $2.522 billion in 2024, and reaching $7.326 billion for all of 2025. In 2026, compute spending in the first half alone already reached $8.945 billion, exceeding the full-year total for 2025.

Moreover, Anthropic has signed long-term compute and infrastructure contracts worth more than $650 billion with eight companies including Amazon, Microsoft, and Nvidia, covering the next decade, of which about 80% are non-cancelable or payable regardless of actual usage.

Soaring infrastructure spending is a dilemma facing the entire AI industry. OpenAI expects to invest about $856 billion in compute and infrastructure from 2026 to 2030; Google's capital expenditure rose to $44.9 billion in the second quarter of 2026, while its free cash flow for the quarter fell to negative $5.9 billion, the first negative turn since its listing. In addition, Meta, Amazon, and others have also seen free cash flow turn negative and capital expenditure guidance raised.

This also reflects that at the current juncture, value distribution across the AI industry chain remains extremely unbalanced. Of every $1 of token-related revenue, the hardware layer accounts for about 70% of industry-wide revenue, the cloud services layer about 20%, and the model and application layer less than 10%. In terms of operating profit, the data are even more extreme: JPMorgan's "China AI Value Chain: An Expanding Pie and the Fight for Value" calculates that in 2026 the hardware segment will account for 93% of industry operating profit, while profit at the model and application layer is close to zero; this situation is expected to improve only by 2030: the hardware profit share will shrink to 30%, cloud services to 15%, and models and applications will rise to 54%.

The latest data show that Anthropic's Q3 single-quarter revenue was about $16.9 billion, and its quarter-over-quarter growth rate has slowed from 144% in Q2 to about 47%. The prospectus also gives a warning about this: recent growth may reflect early adoption and early-market characteristics and may not be sustainable; the company expects the growth rate to decline as the revenue base grows larger, and it may decline faster and more than expected. Among the risk factors, the company directly states that it may be unable to achieve or maintain profitability in the future.

Currently, model-layer companies such as Anthropic are in the middle of the industry chain, where profits are thinnest but capital consumption is heaviest, which also confirms the core logic of the analogy that "AI will become water, electricity, and coal." As infrastructure, they are indispensable, but profit margins are constrained by factors such as regulation, competition, and capital intensity, and the greatest economic value ultimately flows to those who use them, not those who provide them. In the future, it is more likely to be the application layer that truly holds greater profit space.

Whether model companies, including Anthropic, can carve out a sufficiently large share of the pie in the application field remains unknown.

As for the judgment that model-layer profits are thin and value ultimately flows to the application layer, Dario, who is in the middle of it, has a clear understanding. In an interview with Z Finance, he once admitted: "I think opportunities still exist in large quantities at the application layer."

But at the same time, he also raised an objection: model companies release a new version every two or three months, and each upgrade breaks the original capability boundary, suddenly making products that were previously impossible feasible. In Dario's view, this is exactly where the vitality of the API model lies. It does not statically move toward commoditization, but is dynamic and constantly flowing. As long as the boundary of what AI can achieve continues to expand, APIs are the tool for entrepreneurs to seize these new boundaries at the first opportunity.

This also explains why Dario is willing to bear a computing bill of more than $650 billion while still maintaining high-frequency model iteration. Because he knows well that only if the underlying model capabilities continue to iterate can the application layer prosper. What entrepreneurs should think about even more is: "What new things can I build on these models and APIs that others cannot? What truly new ideas can I propose?" Channel: Finance and Economics. This content is published with the author's authorization, and the views represent only the author himself 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[ ]