Anthropic Debuts on Nasdaq, Valuation Nears $2 Trillion
Anthropic debuted on Nasdaq on the evening of November 19, Beijing time, at an offering price of $83 per share. Its indicative opening price was more than 20% higher, implying a valuation approaching $2 trillion. Revenue rose from $41 million in 2023 to $4.6 billion in 2025 and reached $33.24 billion in the first three quarters of 2026. The company also disclosed more than $900 billion in long-term computing and infrastructure purchase commitments, including $70.7 billion in spending over the coming year.
Why read this
SelectedAI score 95
Drawing on the prospectus, the article examines Anthropic’s surging revenue, computing commitments, and investment returns, providing context for assessing how large-model companies will fare in public markets.
Topics
Body · AI translation
The sound of an era turning the page is coming from afar.
On the evening of November 19, Beijing time, Anthropic officially debuted on the Nasdaq, with siblings Dario and Daniela Amodei appearing at the bell-ringing ceremony. Its shares were priced at $83 apiece, while the opening indicative price was more than 20% higher, corresponding to a valuation approaching $2 trillion, or more than RMB 10 trillion.
That unleashed a staggering windfall: each of the seven co-founders saw their net worth rise to about $20 billion, totaling more than $140 billion and collectively propelling them near the top of the Bloomberg Billionaires Index. Early employees and investors also gained an opportunity to cash in. Some finally reaped their rewards, while others were left with regrets—even Masayoshi Son lamented on the eve of the listing, “I really wish I had invested back then.”
It is hard to imagine that this trillion-dollar empire began with a split just six years ago. The Amodei siblings left OpenAI over differences in philosophy, faced rejection after rejection, and ultimately used Claude Code to pry open the door.
Former Baidu employee builds a trillion-dollar company in six years
The Amodei siblings’ departure from OpenAI was highly dramatic.
It was December 2020, and OpenAI was holding an all-hands video conference. After Sam Altman finished reporting on research progress, Dario took over and announced that he, his sister Daniela, and five other colleagues would be leaving to strike out on their own. Immediately afterward, the seven were removed from the meeting, and their work email accounts and computers were disabled.
ChatGPT had not yet been launched.
Dario had previously worked at Baidu. After earning a PhD in biophysics from Princeton University, he took his first job at Baidu, helping develop the Deep Speech 2 speech-recognition system under Andrew Ng. He did not attract much attention at the time, but ultimately joined OpenAI and led the development of GPT-3. His sister Daniela studied English literature at the University of California, Santa Cruz, and later oversaw safety and policy at OpenAI. They criticized OpenAI for moving too quickly toward commercialization without giving safety a corresponding increase in attention.
In 2021, the siblings and their colleagues jointly founded Anthropic, a name meaning “related to humanity.” After completing the first version of Claude, the team remained uncertain about the model’s safety and delayed its release by around six months. By the time Claude made its public debut in March 2023, ChatGPT had already captured all the attention.
The turning point came through an accident.
An engineer casually asked Claude what song was playing on the computer. Unexpectedly, the model wrote a script on its own to identify the song, using a language the engineer was not even familiar with. That internal experiment eventually evolved into Claude Code. Developers could give the model tasks directly through the command line and have it read, modify, and test code.
The enormous computing investments that had once been difficult to cover had finally found a rapidly expanding business model.
The company’s prospectus laid bare its commercial model. Anthropic generated just $41 million in revenue for all of 2023. That figure surged to $4.6 billion in 2025, and then soared to $33.24 billion in the first three quarters of 2026, including $16.96 billion in the third quarter alone.
Even more striking was the quality of its customers. By entering through programming and workflows, Anthropic secured a firm grip on the enterprise market, whose customers have the greatest willingness to pay. As of the end of September this year, it had more than 11,900 customers paying over $100,000 annually, while its net revenue retention rate among existing customers reached an extraordinary 900%.
The outlines of profitability are emerging, but the company’s appetite for computing power is equally astonishing. The prospectus revealed an astronomical bill: to support continued model development, the company has signed more than $900 billion in long-term computing and infrastructure purchase commitments, including as much as $70.7 billion in spending over the next year alone.
History often unfolds with a touch of irony. The people who once left out of concern about excessive commercialization took less than six years to build an even larger enterprise. Such speed has never been seen in the history of human commerce.
At the offering price of $83 per share, each of Anthropic’s seven co-founders holds shares worth approximately $20 billion, for a combined value exceeding $140 billion, placing them near the top of the Bloomberg Billionaires Index.
Employees are also gaining an opportunity to realize their wealth. Anthropic now has more than 5,000 employees, who stand to receive gains of more than $200 billion from the IPO. More than three-quarters of the first 100 employees to join the company are still there today. A career choice made several years ago now carries life-changing weight.
The earliest investors have finally secured returns of more than a hundredfold.
Back in 2021, less than six months after Anthropic was founded, Skype co-founder Jaan Tallinn led its Series A round, with former Google CEO Eric Schmidt and others participating. The round raised about $125 million in total. At the $83 offering price, those shares are worth approximately $40.3 billion, representing a return of more than 300 times the original investment.
Another investment came from Hollywood. Sound Ventures was co-founded by Madonna’s manager Guy Oseary and actor Ashton Kutcher. Having already invested in OpenAI, they obtained consent from both Altman and Amodei before investing about $35 million in Anthropic. They continued adding to their stake, ultimately investing approximately $435 million in total. Based on the stake disclosed before the listing and a rough estimate using the offering price, their shares are worth about $7.5 billion, roughly 17 times their cumulative investment.
Spark Capital also posted a stunning record. As the only venture capital firm named in the prospectus, it invested a cumulative $777 million, and the market value of its holdings soared to approximately $14 billion, for an overall return of about 18 times; its initial $75 million investment alone generated a potential return of nearly 80 times. Menlo Ventures, meanwhile, invested approximately $1 billion in total and received a stake worth about $21 billion, representing a return of roughly 22 times.
Turn the clock back a few years, and behind every early investment was a story of an awkward struggle to attract interest. Investor Anjney Midha still remembers vividly that, while helping Anthropic raise funding, he approached 22 VC firms in one go and was rejected 21 times.
That is precisely why the institutions that later caught up half a beat behind have even more mixed feelings. ICONIQ, now the largest financial-investor shareholder, once passed up the opportunity. It did not return until 2025, when it led the Series F at a lofty valuation of $183 billion, and it continued to increase its investment thereafter. In total, it invested more than $6 billion for a stake worth approximately $30 billion. Sequoia Capital also hesitated because of its bet on OpenAI. Although it ultimately appeared on Anthropic’s lists of Series G and Series H shareholders this year, its entry cost was by then in an entirely different league.
Amazon and Google have reaped even greater gains. By calculation, they invested approximately $13 billion and $13.6 billion, respectively, and the estimated value of each company’s stake has already exceeded $100 billion.
Amid all the excitement, some missed out on an entire mountain of gold.
FTX founder SBF once invested $500 million in Anthropic through Alameda Research. After the company went bankrupt, the related shares were sold off in two batches in 2024 to repay creditors, yielding approximately $1.34 billion. Jane Street, HOF Capital, funds under Fidelity, and other institutions acquired some of them. Had the shares been held until this IPO, they would now be worth approximately $37 billion.
SoftBank founder Masayoshi Son also missed the opportunity.
“I really wish I had invested back then,” Son said ruefully on the eve of the listing. Recalling a meeting with Dario, he said he had greatly admired the founder. SoftBank had also discussed taking a stake in Anthropic, but by then it had begun investing in OpenAI, and Son felt it would be inappropriate to also bet on a rival of its partner.
In just over five years, Anthropic went from struggling to attract any interest to commanding a trillion-dollar market capitalization. Some cashed in their chips by holding firm, while others hesitated and missed the defining opportunity of an era. Yet ringing the opening bell is never the end of the story.
The Watershed
“Never let the company’s valuation exceed its true value, even by one dollar.” Before the listing, Dario said he hoped Anthropic would maintain the same habit in the capital markets: underpromise and overdeliver.
With a trillion-dollar IPO at hand, those words are particularly thought-provoking.
Over the past few years, large-model companies have lived in the greenhouse of the private market. As long as their technology curve was steep enough and their vision grand enough, VCs were willing to keep investing in their future potential. But the rules change upon entering the public market: compute bills, customer retention, and the path to profitability must all withstand ruthless scrutiny.
Current market sentiment offers the clearest annotation. Since the beginning of November, the Nasdaq-100 Index has shed nearly $2 trillion in market capitalization, and the market’s patience with lofty AI valuations and astronomical capital expenditure is wearing thin. It is amid this climate of pervasive alarm that Anthropic has defied the headwinds and stepped onto the bell-ringing podium.
No one can provide the answer in advance. As a trillion-dollar large-model company venturing onto the ice, every candlestick on Anthropic’s post-listing chart will test the thickness of the ice for the entire industry.
Wall Street is not alone in holding its breath and watching the stock. OpenAI may be the most anxious observer. The company secretly filed for an IPO in June this year, with Altman personally driving its restructuring and targeting a 2027 listing. With a valuation of $1.25 trillion, it arguably needs Anthropic to hold steady in the public market more than anyone else.
Across the sea, China’s large-model companies are watching this major test just as closely.
When it comes to IPOs, the Chinese contingent moved earlier. In January this year, Zhipu and MiniMax listed one after another on the Main Board of the Hong Kong Stock Exchange. In June, the Shanghai Stock Exchange issued review guidelines explicitly making the fifth set of STAR Market listing standards applicable to large-model companies, opening the door to high-quality technology companies that have yet to generate revenue at scale.
Reports subsequently emerged that DeepSeek was preparing for a STAR Market IPO and had hired CITIC Securities. Within a matter of months, DeepSeek advanced two funding rounds alongside its listing preparations. The market has even begun to quietly anticipate that, if its market capitalization eventually surpasses RMB 1 trillion, founder Liang Wenfeng’s paper wealth could reshape the rankings of China’s rich list.
The queue for Hong Kong listings continues to grow. Investment community outlet Pedaily previously learned exclusively that Moonshot AI, the company behind Kimi, had notified investors of a restructuring in preparation for a Hong Kong IPO. StepFun, Kling AI, and others have also reportedly developed listing plans.
The growing pains of share-price fluctuations are unavoidable, but they may not be a bad thing for the industry. Bubbles supported by concepts alone must ultimately be cleared away, while technologies truly capable of changing the era have never had reason to fear leaving the greenhouse. Only by withstanding the public market’s most exacting scrutiny can they truly put down roots.
Looking beyond short-term gains and losses, the bell-ringing itself sends a clear message: devoting one’s youth and energy to the most difficult foundational technologies is no longer merely an austere form of idealism; it can also yield respectable and substantial rewards.
For China’s technology sector today, such rewards are worth looking forward to. Technological innovation is, by its nature, a long and solitary process of trial and error. New names on the rich list do more than change the destinies of a select few; they also form silent signposts, giving those who follow the resolve to persevere. The advantages of a new generation of industries may well be taking shape in these choices.
As Liang Wenfeng once said, “When society allows those who pursue hard-core innovation to achieve success and recognition, the collective mindset will change.”
This article is from the WeChat public account “Pedaily” (ID: pedaily2012), written by Wang Lu and published by 36Kr with authorization.
AI Trends