The AI industry is giving rise to a new super IPO.
On June 1, multiple media outlets reported that Anthropic has officially submitted its IPO filing confidentially to the U.S. SEC.
Just three days ago, the company completed a $65 billion Series H round, bringing its post-money valuation to $965 billion (approximately RMB 6.9 trillion)—a figure that has now surpassed OpenAI for the first time.
Anthropic has also become the world's most valuable AI company. If its IPO proceeds smoothly, it could become one of the largest technology IPOs globally in recent years.
Even more astonishing: the company was founded just five years ago, and its founding team consists of only seven people, all of whom came from OpenAI.
Based on the current valuation, the net worth of the seven co-founders is equally impressive. However, due to their relatively low actual ownership stakes (generally below 1%), different institutions use varying calculation methods.
Bloomberg, citing internal equity structures, estimates each person holds approximately 0.83%, with a net worth of about $8 billion (approximately RMB 57 billion), totaling around $56 billion (RMB 380 billion) for the seven individuals.
In a single day, seven billionaires were created, the highest number ever generated by a single company since the inception of the Bloomberg Billionaires Index.
But perhaps the most astonishing thing about Anthropic isn't its valuation—it's that it’s allowing the entire AI industry to see for the first time: AI has begun to take over high-paying labor markets.
$965 billion, approaching the scale of a national-level company
Many people don’t have a clear sense of what $965 billion means. To put it another way, Anthropic’s current valuation has surpassed that of AMD, Intel, Uber, Salesforce, and Qualcomm, and is approaching the combined market value of all technology companies in Germany’s DAX index—despite having been founded only five years ago.
Even more astonishing is the pace of growth. In September 2025, Anthropic’s valuation was only $183 billion; now it stands at $965 billion, surging approximately fivefold in just eight months. If we extend the timeline further, Anthropic’s valuation was only around $6 billion at the beginning of 2024, representing more than a 150-fold increase in just over a year. This may be one of the fastest valuation growths in the history of major tech companies.
On average, Anthropic’s theoretical daily valuation increase is about $2 billion, equivalent to gaining a new unicorn every day.
Why would the capital market dare to offer such an insane price? There's only one answer: revenue.

Anthropic's Funding Journey (2021–2026); Data Source: Crunchbase / Caixin Magazine
Revenue increased 5,400 times over 27 months
What truly supports the $965 billion valuation is not the concept, but Anthropic’s nearly runaway revenue growth.
According to CEO Dario Amodei, Anthropic has grown approximately tenfold each year since earning its first revenue.
Public data shows: ARR was approximately $10 million in 2022, approximately $100 million in 2023, approximately $1 billion by the end of 2024, approximately $7 billion in September 2025, approximately $9 billion in December 2025, approximately $14 billion in February 2026, approximately $30 billion in April 2026, and exceeded $44 billion in May 2026.
What does this mean? Over 27 months, it grew from $870 million to $44 billion, an increase of approximately 5,400 times. This pace far exceeds that of the traditional SaaS era and even surpasses the growth records of many internet companies.

Anthropic ARR Growth Trend (2022–2026); Data Source: Disclosed by CEO Dario Amodei / China Fund News
But Anthropic’s most core growth engine is not its chatbot, but Claude Code. This AI-powered programming product, launched by Anthropic, has achieved over $1 billion in annual recurring revenue (ARR) in just six months. By comparison, most SaaS companies in history have taken over a decade to reach $1 billion in ARR. Claude Code did it in half a year.
Even more striking, approximately 4% of all open-source code on GitHub worldwide has been generated by Claude—meaning that for every 25 new lines of code added globally, one comes from Anthropic.
In the SWE-bench Verified evaluation, Claude Code scored 93.9%, significantly outperforming OpenAI Codex’s 80.8%. What the capital market is truly betting on is not chatbots, but AI beginning to take over programming tasks.

Anthropic Claude interface; image source: Anthropic official website
1,000 companies, each giving it $1 million per year
The truly alarming aspect of Anthropic is not the consumer market, but the enterprise market.
Public data shows that Anthropic has over 1,000 enterprise customers paying more than $1 million annually. If calculated simply, the theoretical annual revenue from these customers alone exceeds $1 billion.
More importantly, these are not ordinary companies—they are Cursor, Notion, Zoom, Slack, financial institutions, and global software companies. Over the past year, Anthropic’s enterprise market share increased from approximately 10% to over 65%.
A major shift is occurring in the AI industry: it’s no longer individuals spending money on AI, but companies beginning to allocate budgets of millions or tens of millions of dollars to procure AI labor.
Liu Run, founder of Runmi Consulting, summed it up plainly: OpenAI sells emotional value, while Anthropic sells productivity.
OpenAI enables chatting, drawing, and video creation, while Anthropic helps businesses automatically write code, approve processes, analyze data, migrate systems, and manage risk. The B2B market is far larger than the B2C market.
Anthropic Claude interface; image source: Anthropic's official website. OpenAI has nurtured its biggest rival.

Dario Amodei (CEO of Anthropic); Image credit: World Economic Forum / TechCrunch
Anthropic's story is essentially a split within OpenAI.
Back in 2020, when GPT-3 had just been released and OpenAI was at the height of its momentum, Dario Amodei, founder of Anthropic, was then the Vice President of Research at OpenAI, overseeing the core development of GPT-2 and GPT-3.
OpenAI once allocated 50%-60% of its total computing power to him.可以说, GPT-3's development was heavily reliant on Dario, making him one of the most pivotal figures.
But as OpenAI became increasingly commercialized, internal divisions emerged. According to multiple media reports, OpenAI internally discussed whether it might sell AGI capabilities to global nuclear powers—a prospect Dario found utterly unacceptable. In his view, AGI should never become a tool for endless profit-seeking.
Ultimately, at the end of 2020, Dario decided to leave OpenAI—and he didn’t go alone. He took five key OpenAI researchers with him: Tom Brown, lead author of GPT-3; Jared Kaplan, author of the Scaling Laws paper; Chris Olah, expert in AI interpretability; Jack Clark, head of OpenAI policy; and Sam McCandlish, a researcher at OpenAI—along with his sister Daniela Amodei, making a total of seven people.
In 2021, Anthropic was founded. The company's name, "Anthropic," means "related to humans." No one anticipated that five years later, the company would be valued at $965 billion.
And OpenAI, for the first time, faces its most dangerous rival.
Why can seven people become wealthy together?
One of the most counterintuitive things about Anthropic is equal voting rights: the seven founders each hold nearly identical shares (currently less than 1% each).
Many people back then thought this design would fail, because the Silicon Valley startup logic has always been that there must be an absolute core founder. But Dario believed that Anthropic wasn’t built around any one person—it was built around a mission. As a result, this “seemingly silly” decision became Anthropic’s strongest stabilizer.
Of OpenAI’s original 11 core founders, most have since left, while all seven of Anthropic’s co-founders remain after five years.

Comparison of net worth of AI company founders/core figures; Data source: Forbes / Bloomberg Billionaires Index
According to publicly available data, Anthropic’s two-year employee retention rate is approximately 80%, higher than OpenAI’s 67%. In a rare occurrence, many CTOs from top companies are willing to take a demotion and join Anthropic as regular engineers, because in today’s AI industry, what’s most valuable is no longer job title, but equity.
Who else made a killing?
Behind this round of funding for Anthropic is a long list of winners.
Most striking is Google: in 2023, Google acquired approximately 10% of Anthropic for just $300 million. Based on a $96.5 billion valuation, this stake is worth about $96.5 billion—more than 300 times its original value in three years.
Amazon has also made a massive bet, investing over $13 billion in total. The reason is simple: Anthropic needs computing power, and Amazon needs AI traffic—the two are now deeply intertwined.
Even more astonishing, Anthropic is now paying SpaceX approximately $1.25 billion per month in computing costs, totaling nearly $45 billion over the three-year contract.
In other words, Anthropic isn't just creating AI billionaires—it's also aggressively fueling GPU manufacturers, cloud providers, storage companies, data centers, power utilities, and the entire AI supply chain. The AI industry has entered a true arms race.
The fastest wealth creation in AI history has just begun.
But problems have also begun to arise.
First, most of Anthropic's current wealth is paper wealth, as the company has not yet officially gone public, and many shares cannot be liquidated.
Second, AI is extremely costly. Public data shows that Anthropic's annual compute costs are approximately $15 billion, with compute costs still accounting for more than 50% of revenue.
Third, a $965 billion valuation is already close to "perfect expectations," meaning that even a slight slowdown in growth would create significant valuation pressure.
Even so, capital markets continue to make bold bets, as everyone is beginning to realize that the most profitable phase of the AI industry may have just begun.
In the past, the internet earned money through advertising; mobile internet earned money through traffic; but AI truly aims to earn money from human labor. Tasks such as coding, customer service, analysis, research, operations, and sales—many of which once required human effort—are increasingly being absorbed by large models.
Five years ago, Anthropic’s seven founders left en masse due to dissatisfaction with OpenAI’s increasingly aggressive commercialization path. Five years later, the company they founded has surpassed OpenAI in valuation.
But what’s most alarming about Anthropic may not be the $965 billion—it’s that the world has, for the first time, truly seen: AI has begun to take over high-paying labor markets.
This article is from the WeChat public account "Pencil News" (ID: pencilnews), authored by Pencil News.
