Anthropic Debuts on Nasdaq, Valuation Briefly Nears $2 Trillion at the Open
Anthropic listed on the Nasdaq under the ticker ANT, with its first trade at $111.50, up 34.34% from the $83 offering price, briefly bringing its fully diluted valuation close to $2 trillion. The IPO's public offering totaled approximately $81.3 billion; as of publication, the share price had retreated to $94.62, giving the company a market capitalization of about $1.24 trillion. The prospectus also disclosed substantial spending on computing power and infrastructure, as well as the company's use of Claude to assist in developing its next-generation models.
Why read this
SelectedAI score 87
The article provides a comprehensive overview of Anthropic's listing through its offering size, opening performance, and computing commitments, offering insight into how capital markets value frontier AI model companies.
Topics
Body · AI translation
On November 19 local time, Anthropic, the company behind Claude, listed on Nasdaq under the ticker ANT. Its first trade was executed at $111.50 per share, up 34.34% from its $83 offering price.
On a fully diluted basis, Anthropic’s valuation rose from about $1.5 trillion at the time of the offering to roughly $2 trillion. The share price later pulled back. As of publication, ANT was trading at $94.62, giving the company a market capitalization of about $1.24 trillion.
From the moment of its debut, it had already reached heights many companies could never attain in their lifetimes.
When Elon Musk’s SpaceX went public this June, we had only just used the phrase “the largest IPO in history.” Who could have imagined that the record would last only five months? Anthropic’s IPO raised about $81.3 billion in its public offering.
In the end, Anthropic beat OpenAI to the punch, becoming the first independent, publicly listed frontier AI model company in the United States.
That also made Anthropic, to some extent, the face of AI. In New York’s Times Square, protesters began arriving early in the morning. Anthropic’s name appeared on the screens of the Nasdaq building, while a red banner below read, “Freeze AI.”
The day before the opening-bell ceremony, Hugging Face co-founder Thomas Wolf, Turing Award winner Judea Pearl and others signed an open statement. Their criticism targeted precisely the reputation Anthropic values most. In their words: “Anthropic is not a safe AI company.”
The next morning, Dario Amodei and his sister Daniela stood together before Nasdaq’s opening bell. During her remarks, Daniela made a point of addressing those who had misgivings about AI, including people who had never even used Claude. “This technology needs you,” she said.
The largest IPO in history is born
Going by the prospectus filed nine days earlier, SpaceX would still have held the title of the largest IPO in history.
When its roadshow began on November 10, Anthropic planned to sell 825 million shares at $70 to $78 each. Even at the top of that range, the offering would have raised only $64.35 billion, more than $10 billion short of SpaceX’s $75 billion record. But just two days later, investors were seeking roughly six times the number of shares originally on offer.
The siblings split up to meet with investors during those days. Dario spent an entire day at JPMorgan, while Daniela and Chief Financial Officer Rao went to Goldman Sachs to meet with its wealth-management clients. By the evening of November 18, Anthropic had increased the offering to 980 million shares and raised the price to $83 per share.
Multiply the two figures and the result is $81.34 billion, pushing SpaceX into second place.
JPMorgan arranged a livestream for clients for the listing and even prepared T-shirts and laptop sleeves. Anthropic itself did not organize so much as a single celebration party.
Times Square was bustling on the day of the bell-ringing ceremony. At 9 a.m., protesters unfurled a banner as someone played drums behind them. A street performer dressed as Minnie Mouse also stopped to watch, and there were more members of the media taking photos than there were protesters.
Inside the building, Daniela took the stage first. She said some people might find Anthropic a “strange” company. She then added that quite a few people had, in fact, already said so.
Dario took the stage next and spoke about their father. Their father had suffered from hepatitis C, and just a few years after his death, new drugs offering far better treatment became available. Dario said he had originally gone into biological research because he wanted to help patients gain access to new medicines sooner.
Now he believes AI can help with that as well. Scientists are already using Claude for tasks ranging from studying proteins and designing drugs to organizing clinical trials.
At 9:30 a.m. New York time, the countdown ended, and the siblings rang the opening bell together with their colleagues. Alongside Rao, who had accompanied Daniela to meet investors, was co-founder Jack Clark. Ten years ago, Clark was still a technology reporter at Bloomberg. He later left to join OpenAI before departing with the Amodei siblings to found Anthropic.
The bell had rung, but those watching ANT’s share price on their screens were still waiting. Investors who subscribed the night before had bought shares at $83 each; now everyone wanted to know how high the opening price would go.
At 10:10 a.m., the opening reference price was set at $104. Sixteen minutes later, it rose to $105. It was not until 11:48:24 a.m. New York time that the first trade was executed at $111.50 per share, up 34.34% from the offering price.
Those who hit the jackpot
The three people who rang the bell onstage—Dario, his sister Daniela and co-founder Jack Clark—each sold exactly the same number of shares in the IPO. Each sold about 3.38 million shares, worth $280 million at the $83 offering price, for a combined total of $840 million.
What they sold was only a tiny fraction of their holdings. At the opening price, Dario’s remaining roughly 237 million shares were worth $26.45 billion. His sister held slightly more than her brother, about 238 million shares worth $26.54 billion. Clark held slightly more than Daniela, with approximately 239 million shares worth $26.64 billion.
Big technology companies also profited. Amazon and Google together owned nearly 30% of the company. After the offering and concurrent private placement were completed, Amazon held a 14.6% stake and Google 13.4%, exactly 28% combined. At the opening price of $111.50 per share, their holdings were worth approximately $413.8 billion in total.
Of course, what most workers probably want to know is this: If they had taken a job at Anthropic back then, how much would they be getting today?
Indeed, even an administrative employee at Anthropic could have vaulted into the ranks of dollar millionaires.
In 2024, Anthropic was hiring administrative staff in San Francisco to manage executives’ schedules, organize meetings, and process expenses. The annual cash salary ranged from $140,000 to $190,000, plus equity. The options granted to employees that year had an average exercise price of $1.30 per share.
Suppose an administrative employee had options to buy 12,500 shares at that price. After deducting the cost of exercising them, the employee’s pre-tax paper gain at the offering price would already have been about $1.02 million. At the opening price, that gain rose to approximately $1.38 million.
Chief Financial Officer Krishna Rao, who joined the same year, also received options with an exercise price of $1.30 per share—14 million shares in total. For that initial grant alone, the gap between the offering price and exercise price amounted to $1.14 billion. At the opening price, it widened to $1.543 billion.
Before the listing, research firm Sacra ran the numbers using a valuation of $1.35 trillion. Including both current and former employees, it estimated that around 2,800 people held vested equity worth more than $1 million. Private-share trading platform Hill excluded all the founders and still estimated that 598 employees might each hold equity worth more than $100 million.
Interestingly, people in Los Angeles are already targeting this newly rich cohort. The seller of a five-bedroom home listed for nearly $30 million said Anthropic shares could be used to pay for it.
Speaking of which, while someone in Hollywood wants to exchange a house for shares, others sealed their investment in their own living room years ago.
In 2023, Dario visited the Beverly Hills home of Madonna’s manager, Guy Oseary. Actor Ashton Kutcher was also there. After talking for several hours, Oseary and Kutcher decided to invest in Anthropic through Sound Ventures, the fund they co-founded.
There was also a small complication: the fund had already invested in OpenAI, so Kutcher first gave Altman a heads-up. The fund ultimately invested $435 million in Anthropic over several rounds. Based on its estimated ownership stake, those shares were worth about $10 billion at the opening price—23 times the original investment.
Some reaped a rich harvest; others were left kicking themselves.
One investor who got in even earlier than the two Hollywood figures was SBF, founder of the cryptocurrency exchange FTX.
In 2021, just after Anthropic was founded, he invested $500 million through Alameda. After FTX went bankrupt, its liquidation team sold all those shares in two transactions in 2024, recovering a total of $1.336 billion to repay customers and other creditors. At the $83 offering price, the shares would have been worth approximately $37 billion. Had they waited until trading opened, they would have been worth about $49.7 billion.
A Future Priced at $2 Trillion
Software company Grafana Labs has a new rule: when employees need access to the most powerful AI models, they get it for only 24 hours at a time. The rule arose after a data analyst forgot to turn off access after using one; by the time anyone noticed, tens of thousands of dollars had already been spent.
Sometimes, however, the money is well spent. One team paid tens of thousands of dollars over a month to use AI to speed up one of the company’s services. Once the work was done, the change alone saved a six-figure sum in Amazon cloud-service fees. Grafana’s chief technology officer told the media that the company now spends tens of millions of dollars a year on AI usage, with roughly two-thirds going to Claude.
And plenty of companies are willing to spend that kind of money.
As of the end of September this year, around 250 customers had each spent at least $10 million on Anthropic over the preceding 12 months.
Anthropic generated $1.4 billion in revenue in the third quarter of last year. In the same quarter this year, that figure had reached $17 billion.
Why would a company capable of doing $17 billion in business in a single quarter still need to raise so much money at once? The answer becomes clear in the pages of the prospectus detailing its spending.
Compute and infrastructure alone cost nearly $9.9 billion in the third quarter of this year. Even greater expenses lie ahead. Anthropic has already signed approximately $900.9 billion in procurement and lease contracts, nearly all related to compute and infrastructure. In the first 12 months after the end of September this year, it expects to pay $70.7 billion.
Among the recipients of that money is Google, which just spent money buying shares. In this placement, Google subscribed for $10 billion of Anthropic stock. Conversely, under the cloud-services agreement between the two companies, Anthropic must spend at least about $121.3 billion with Google from April this year through July 2033. If it orders more compute than it ultimately uses, it will also have to make up the difference under the agreement.
For Google, the more popular Claude becomes, the better.
Some of that computing power is also being used by Claude to develop the next generation of Claude. According to Anthropic, more than 90% of the code added to the company’s codebase in September this year was written by Claude. It is already helping researchers run experiments and prepare training data, and it even plays a role in safety testing new models.
That brings us back to the public statement issued the day before the opening bell.
Thomas Wolf, Judea Pearl, and others signed that public statement the day before the opening bell because they were worried about precisely this: Claude becoming increasingly proficient at such work. Anthropic itself has warned that once AI can improve itself generation after generation, humans may gradually lose control of it. Yet in its prospectus, Anthropic told investors that one of its major strengths is that each generation of Claude can help researchers build the next generation more quickly.
Dario also addressed this concern before ringing the bell. He said the company would spend however long it took to become confident that a new model was safe and remained under human control before training and releasing it; business pressure could not be allowed to rush the process. In his view, going public actually made this a little easier. With more money on hand, the company would be in a stronger position to slow down.
The day before the opening bell, at an event in New York, the host put the same question to Altman: In your view, is what Anthropic is doing now safe?
“I don’t know,” Altman said. “I don’t think they really understand what we’re doing either.”
A few minutes later, the host asked about the valuations of the two companies. Anthropic was set to price its offering that night—how would he compare OpenAI with it? Altman did not put a number on his own company. He said valuations always rise and fall, then turned to the AI industry as a whole.
Altman said OpenAI also planned to go public.
Five years ago, the Amodei siblings and several former colleagues left OpenAI and founded Anthropic. Five years later, they rang the bell first. So far, the siblings’ story remains a wish-fulfillment fantasy come true.
AI Trends