The Giant in Large AI Models Initiates Process for U.S. Stock Market Listing

Author | Yezi

Tap to listen to the podcast version of this article

A Trillion-Dollar Giant Prepares for Public Debut

On June 1, AI behemoth Anthropic confidentially submitted its S-1 filing to the U.S. Securities and Exchange Commission (SEC), marking the formal commencement of the IPO process for this super unicorn, valued at nearly one trillion U.S. dollars.

Just days prior, Anthropic had secured a staggering Series H funding round of 65 billion U.S. dollars, propelling its post-money valuation to 965 billion U.S. dollars, approximately 6.9 trillion yuan. This valuation not only surpasses that of OpenAI (valued at 852 billion U.S. dollars) but also equals the combined market capitalization of three Alibabas.

Consequently, Anthropic's seven co-founders have collectively ascended to the top 500 of Bloomberg's Global Billionaires Index, each holding shares worth roughly 8 billion U.S. dollars, or about 57 billion yuan. This remarkable achievement sets a record for the highest number of new billionaires from a single company in a single day since the inception of the Bloomberg Billionaires Index.

Behind Anthropic's meteoric rise in valuation lies robust demand in the enterprise market. Its annual recurring revenue has skyrocketed from 9 billion U.S. dollars at the end of 2025 to over 40 billion U.S. dollars, with the number of large clients paying over 1 million U.S. dollars annually growing exponentially.

So, what has this team done correctly? What commercial rationale and potential risks underlie this AI wealth phenomenon?

Confidential IPO Filings Create 7 Billionaires Overnight

Anthropic's confidential filing was undoubtedly a meticulously planned move.

The industry widely anticipates that Anthropic's competitors, OpenAI and SpaceX (now merged with xAI), are also gearing up for listings around the same time, keeping global capital on high alert.

Confidential filings are a unique mechanism for U.S. IPOs: after submitting materials, the SEC conducts a closed-door review, and the complete prospectus is only made public shortly before the official listing. The advantage is clear—if the IPO plan is terminated for any reason, the company will not divulge large amounts of core internal data.

In fact, the confidence underpinning Anthropic's IPO lies in its rapidly escalating valuation.

In early 2024, Anthropic's valuation hovered around 6 billion U.S. dollars; by September 2025, following the Series F funding, it surged to 183 billion U.S. dollars; in February of this year, after the Series G funding, the valuation climbed to 380 billion U.S. dollars; and by the end of May, post-Series H funding, it soared to 965 billion U.S. dollars.

In just over a year, the valuation has grown more than 150-fold, theoretically adding nearly 2 billion U.S. dollars in value per day. This valuation not only surpasses OpenAI (852 billion U.S. dollars) but also equals the combined market capitalization of three Alibabas. As of the June 1 U.S. stock market close, Alibaba's market cap stood at 300.851 billion U.S. dollars.

Amid this capital frenzy, the wealth of Anthropic's seven co-founders has surged. The seven are Dario Amodei and his sister Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish, and Christopher Olah.

Unlike many Silicon Valley companies that emphasize a single core founder, Anthropic adopted a distinctive "one share, one vote" model. The seven founders hold nearly identical shares, each with a stake of less than 1% (estimated at about 0.83% by Bloomberg), but given the staggering valuation, each has a paper fortune of approximately 8 billion U.S. dollars, or about 57 billion yuan.

This equity structure has also ensured a stable senior management team. Five years on, none of the seven co-founders have departed, and the executive retention rate far exceeds that of many comparable companies. Clearly, a stable and united core team has become a pivotal factor in the high valuations offered by capital.

4-Year Growth of 4,400 Times: AI Secures Major Enterprise Contracts

Capital is always discerning. The near-one-trillion-dollar valuation is justified by Anthropic's monetization capabilities. Its annual recurring revenue (ARR) growth curve supports its massive market cap.

According to company disclosures, its ARR was about 10 million U.S. dollars in 2022; it rose to 100 million U.S. dollars in 2023; surpassed 1 billion U.S. dollars by the end of 2024; and then began exponential growth: reaching 7 billion U.S. dollars in September 2025, 14 billion U.S. dollars in February 2026, and exceeding 44 billion U.S. dollars by May 2026.

In just over four years, it has grown about 4,400-fold. This efficiency far surpasses traditional SaaS (Software as a Service) business models.

Why such high earnings? The underlying rationale is that Anthropic has not focused on C-end users for chatting, drawing, or emotional satisfaction but has instead targeted the more lucrative B-end (enterprise) market.

Take its core AI programming product, Claude Code, as an example. Within just six months of launch, its ARR exceeded 1 billion U.S. dollars. In various programming benchmark tests, the Claude series models have performed exceptionally well, with about 4% of global GitHub open-source code now showing traces of Claude's involvement.

Moreover, Anthropic's tools have demonstrated high usability in complex enterprise logic chains, such as sales approvals, data analysis, and risk control judgments.

Public data indicates that Anthropic has over 1,000 enterprise clients paying more than 1 million U.S. dollars annually, including well-known software companies like Zoom, Slack, and Notion, as well as numerous financial institutions. Over the past year, its market share in enterprise AI spending has surged from about 10% to over 65%.

This underscores a trend: companies are allocating millions or even tens of millions of U.S. dollars to procure AI solutions to replace traditional high-paying roles. What capital is truly betting on is AI's large-scale takeover of the high-wage labor market.

Core Team Starts Anew, Burning Money Conceals Potential Risks

Anthropic's founders hailed from their former rival, OpenAI.

Anthropic's CEO, Dario Amodei, was previously OpenAI's Vice President of Research, leading the development of early core large models. However, by the end of 2020, due to serious disagreements over the company's overly commercialized direction, Dario departed and took several key researchers, including the lead author of GPT-3, to found Anthropic in 2021.

To avoid repeating past mistakes, Anthropic implemented a unique corporate governance structure. It is registered as a "Public Benefit Corporation (PBC)," with true control resting in a trust. The board must consider "the benefits to humanity" rather than just shareholder interests.

This has led to a peculiar situation: although Google invested 300 million U.S. dollars in 2023 to acquire about a 10% stake, and Amazon has invested over 13 billion U.S. dollars in total, neither tech giant has secured a board seat.

However, the giants have not lost out. Take Google as an example: its 300 million U.S. dollar investment is now worth nearly 96.5 billion U.S. dollars based on the current valuation, a more than 300-fold increase in three years.

Nevertheless, amid this capital frenzy, potential risks persist.

As model capabilities iterate, Anthropic's demand for GPUs and accelerator clusters has surged. Public data indicates its annual computing costs have reached about 15 billion U.S. dollars, accounting for over 50% of its revenue.

To secure sufficient computing resources, Anthropic has not only signed major agreements with Google and Amazon but even pays SpaceX about 1.25 billion U.S. dollars monthly for computing power, with a three-year contract totaling nearly 45 billion U.S. dollars. Most of its earnings are reinvested into computing power and infrastructure.

More critically, the 965 billion U.S. dollar valuation means capital markets have set extremely high expectations. If future revenue growth slows slightly or B-end clients' willingness to pay fluctuates, this valuation, propped up by high expectations, will face significant downward pressure.

Fortunately, the founding team remains clear-headed about this. CEO Dario has repeatedly written warnings about the risks of wealth concentration in the AI industry and stated his willingness to forgo massive wealth. The seven co-founders have also publicly committed to donating 80% of their personal fortunes.

References:

"650 Billion! The Giant in Large AI Models Files for IPO, Valuation Exceeds Three Alibabas," Zhidx;

"Anthropic's Seven Co-Founders Each Reach 8 Billion U.S. Dollars in Wealth, Setting a Bloomberg Billionaires Index Record for Most New Billionaires in a Single Day," Wall Street See;

"The Birth of Seven 50 Billion U.S. Dollar Billionaires, Racing Toward a 1 Trillion U.S. Dollar IPO," Pencil News;

"The 'Giant' Announces: It Has Secretly Submitted IPO Filing Materials," CCTV Finance;

"Anthropic Files Prospectus, Targeting a 1 Trillion U.S. Dollar Market Cap," Jiqizhixin;

"Surpassing OpenAI, Anthropic's Valuation Nears 1 Trillion U.S. Dollars," Beijing Business Today.

THE END

Copyright and Disclaimer

1. Content Copyright: Except for quoted public data, policies, and cases, all content is original. Professional data is sourced from authorized databases and government websites, while cases are compiled from real events.

2. Image Licensing: Some images are proprietary or officially licensed, and others are AI-generated. For network images with unclear copyright, ownership remains with the original authors, and any infringement will be removed upon notification.

3. Republishing Rules: Unauthorized republishing is prohibited. Republishing requires retaining the complete source and author information.

4. Liability Disclaimer: This article is a commercial figure observation and industry commentary compiled by the author based on public information. The content is for reference only and does not constitute professional advice. Risks arising from use are borne by the user. Xinzhidongneng reserves the final interpretation rights to this article.