Anthropic goes public, 10 facts worth paying attention to
outside the page
2026.11.20 11:39
· From Guangdong
Full text 5357 words
Text | Outside the page, author | Publisher
A company went public, raising US$81.3 billion, and its market value exceeded US$2 trillion on the first day.
There is no doubt that Anthropic has set a new IPO record in the U.S. capital market.
But what’s noteworthy about this IPO is far more than just a group of millionaires or how many financing records a company has set.
It is more like a concentrated display. The technological value, private wealth, capital commitments and business risks accumulated by the AI industry for many years are being put on the table on the same day.
What’s changing behind this and what hasn’t?
1. AI companies start to fight for capital
Anthropic's public offering of 980 million shares at an issue price of US$83 raised US$81.34 billion. In addition, the company also arranged a simultaneous private placement of US$26 billion.
The two items combined amounted to more than 107 billion U.S. dollars.
What is this concept?
This is a large-scale capital mobilization around AI infrastructure.
What's more noteworthy is that Anthropic is not a startup without revenue. Full-year revenue in 2025 has reached US$4.585 billion, and in the first nine months of 2026 it has reached US$33.242 billion.
Why would a company with rapidly growing revenue raise capital on such a large scale?
The answer is that competing on cutting-edge models requires locking in computing power, chips, and data center resources for the next few years in advance. For Anthropic, financing is not just about expanding its business, but also buying tickets for the next stage of competition.
IPOs of AI companies are changing from financing business expansion to reserving capital for a long-term technology competition.
2. An IPO creates a group of new rich people
When Anthropic goes public, the most direct beneficiaries are not just the founders, but also early employees and investors.
The Wall Street Journal estimates based on the prospectus documents that the restricted stock held by employees is worth about $88.7 billion, and the options after deducting exercise costs are worth about $140 billion. The total amount exceeds US$230 billion and involves approximately 5,300 employees.
According to this estimate, the average equity value per employee is more than $44 million. Of course, this is just a book average. The actual distribution is highly uneven, and it does not mean that everyone can cash out immediately.
But it still reveals a surprising fact: the wealth created by AI is no longer just a game for a few founders and venture capitalists.
A more interesting detail is that someone in Los Angeles is already willing to accept Anthropic stock as payment for a nearly $30 million mansion.
This means that the private equity of AI companies is becoming a certificate of wealth that can be exchanged for real estate in the real world.
3. The market not only wants to buy new stocks, but also pick up old stocks.
This may be the most overlooked design of Anthropic’s IPO.
Traditional IPOs usually set up insider lock-up periods to prevent founders and early shareholders from selling large amounts of stock immediately after the listing.
Anthropic took a different approach.
According to the prospectus, up to about 316 million shares may be eligible for sale from the first trading day after listing. Up to another 3.3 billion shares may be eligible for sale in stages before the end of the 180-day lock-up period for ordinary shareholders. Longer restrictions apply to founders and some major shareholders.
This means that the market is facing much more supply than the 980 million IPO shares issued.
Of course, obtaining the qualification for sale does not mean actual reduction of holdings, nor does it mean that all these stocks will enter the market on the first day.
But it raises a question that the AI private equity market has rarely had to face in the past. When a company that has accumulated a huge amount of private equity over several years suddenly enters the public market, how much liquidity is needed to undertake it?
Anthropic's share price on the first day fell from its first transaction price of $111.50 to $97.17 at the end of the day.
This trend alone cannot prove that supply pressure has overwhelmed the market, but it is enough to remind investors that how high a company's valuation is and whether the market can digest its shares are two different questions.
This is where I think the market tends to underestimate. What investors see is a rush to buy, but what really needs to be observed is the supply of stocks in the coming months and years.
4. All the money in AI ends up going into computing power.
The money raised by Anthropic will not stay in the bank account for a long time.
There is an even more astonishing figure in the prospectus. As of the end of September 2026 and thereafter, the company has signed commitments for purchasing, leasing and other computing power and infrastructure, totaling approximately US$900.9 billion.
Of this amount, approximately US$447.5 billion is expected to be incurred before the end of 2030, and approximately US$69.8 billion is expected to be paid within the next 12 months.
Of course, the $900.9 billion is not a bill that will be cashed immediately. Some of the contracts have termination rights or preconditions, and the specific payment also depends on the delivery and use of computing power.
But this number is enough to show that Anthropic is providing advance planning for the next ten years of computing power.
Google and Amazon are important players.
Google is both a shareholder of Anthropic and a cloud service provider, and Amazon also provides both capital and computing infrastructure.
This forms an interesting industrial chain. The capital market finances model companies, which use money to purchase computing power. Cloud manufacturers continue to build data centers after receiving income, and chip manufacturers and energy suppliers benefit from this.
The AI industry is forming a cycle driven by capital, models, computing power and infrastructure.
This cycle is supported by real business needs, but there is also a point that requires vigilance. Every company in the industry chain is betting to varying degrees that the next company can continue to grow.
If model demand continues to expand, the cycle will continue to intensify. If demand growth is less than expected, computing power contracts, infrastructure investment, and capital returns may be under pressure at the same time.
This is the real risk after the capitalization of the AI industry.
5. Having income does not mean there is a return.
Anthropic's financial numbers are interesting.
In the first nine months of 2026, the company's revenue was US$33.242 billion, a year-on-year increase of more than 12 times; during the same period, the company's operating loss was US$1.26 billion.
But the net loss reached US$104.039 billion.
Looking at this number, it's easy to see that the more Anthropic makes, the more it loses.
This is not the case.
Of this amount, approximately US$94.3 billion came from changes in the fair value of financial instruments such as convertible notes, derivative liabilities and SAFEs. Such projects will be affected by factors such as an increase in the company's valuation, and are not equivalent to the company burning so much cash in actual operations.
There is another noteworthy detail. Anthropic disclosed that adjusted operating profit for the first nine months of 2026 was $4.1 billion, which is the opposite direction of operating losses calculated under accounting standards. This shows that the company's profit performance varies greatly under different calibers.
The question is, with revenue growing so fast, can we generate stable profits and cash flow in the future?
After all, $1.881 billion in the first nine months? No, it’s US$18.808 billion in computing power and infrastructure expenses, which already accounted for about 57% of revenue during the same period.
The stronger the model capabilities and the more customers use it, the revenue may increase, but the demand for computing power is also expanding at the same time.
This divides the AI bubble problem into two levels: Is there a real demand for AI? Anthropic has given quite strong proof.
But can AI achieve returns commensurate with capital investment? This question has not yet been fully answered.
Real industrial demand can support a company's growth, but it cannot automatically justify its current valuation.
This is also the most important stress test for the entire AI industry when Anthropic goes public.
6. Safety ideal meets the test of capital
There are two scenes worth watching together at Anthropic’s listing scene.
Upstairs, Dario Amodei and his sister Daniela rang the Nasdaq opening bell.
Downstairs, protesters held banners reading “Freeze AI.”
The company has always regarded AI safety as its core mission. It even adopts the organizational form of a public benefit company (PBC), which requires the board of directors to strike a balance between the economic interests of shareholders, the interests of affected groups and the long-term public interest.
Anthropic also established a long-term interest trust that gave it the power to elect and remove a majority of its board of directors.
This is a rather special set of corporate governance arrangements. But after going public, it must also face another set of rules in the capital market: growth, competition, profits and shareholder returns.
Claude needs to continue to get stronger to cope with competitors such as OpenAI and Google. The company needs to continue to expand its revenue to support increasingly large investments in computing power.
A contradiction arises. Anthropic hopes to remain cautious about the development of AI, but the business competition it faces requires that it not slow down easily.
That doesn't mean the safety mission will necessarily give way to growth after going public. PBCs and long-term benefit trusts are precisely the institutional arrangements through which companies try to protect their mission.
However, whether the system can withstand the pressure of market competition and capital returns ultimately requires time to test.
On the day of ringing the bell, the two voices upstairs and downstairs just constitute a real situation in the AI era. Capital wants technology to run faster, while society begins to ask who should decide how fast it runs.
7. Claude starts making Claude
There is another number in Anthropic’s prospectus. In September 2026, more than 90% of the code contributions in the company's code base were completed by Claude.
Moreover, Claude doesn't just help engineers write code. It has entered the company's research process and is involved in experiments, training data preparation and model safety testing.
Anthropic also disclosed that from January to September 2026, the value of the company's internal use of Claude increased by more than 250 times based on public prices; in April 2026, this monthly value exceeded the company's personnel expenses excluding equity incentives for the first time in that month.
This means that Claude is not only a product for external sales, but also becoming Anthropic's own production tool.
Models help employees improve efficiency, and employees develop next-generation models with higher efficiency. The new models further improve R&D efficiency.
This can create a reinforcing feedback loop.
It is different from the efficiency improvement of ordinary software companies. If AI can continue to participate in AI R&D, the competitive advantage of model companies may shift from the lead in a certain generation of models to the ability to continuously iterate models.
Of course, 90% code contribution does not mean that 90% of the research and development work has been completed automatically. Anthropic also made it clear that designing and building successive generations of cutting-edge models completely autonomously still requires meaningful human supervision and understanding.
But one thing has changed. AI is no longer just a product manufactured by technology companies and sold to customers. It has begun to become a tool for technology companies to create the next generation of products.
This may be one of the most noteworthy productivity changes in the AI industry.
8. FTX missed Anthropic’s craziest rise
Another dramatic character in Anthropic’s wealth story is the bankrupt cryptocurrency exchange FTX.
In 2021, FTX founder Sam Bankman-Fried (SBF) invested US$500 million in Anthropic through his Alameda Research. At that time, the company had just been established and its valuation was only about US$4 billion.
Later, FTX went bankrupt and the liquidation team needed to repay customers and creditors.
In 2024, the liquidation team will sell all Anthropic shares and recover approximately US$1.336 billion.
The deal looked quite successful at the time: $500 million turned into $1.336 billion, a return on investment of more than 1.6 times.
However, based on Anthropic’s IPO valuation of approximately US$1.5 trillion, the corresponding value of these shares at that time was close to US$37 billion.
For the same investment, the difference was nearly US$36 billion.
Of course, today’s prices cannot be used to criticize the liquidation decision at that time. FTX has gone bankrupt, and the task of the liquidation team is to recover as much money as possible, rather than continue to bear the long-term investment risks of an AI company on behalf of creditors.
But this story reveals a very realistic rule of venture capital. The most profitable investments often need to survive a long period of uncertainty, and the time when they are most needed to be cashed out may also be the time when the future value has not yet fully emerged.
Anthropic’s success has not only created a group of new rich people, but also left behind a group of investors who missed out on greater returns because of early exit.
9. Hollywood is also betting on AI
In 2023, Dario Amodei visited a house in Beverly Hills.
The owner of the house was Madonna's manager Guy O'Sealy, and actor Ashton Kutcher was also present.
Several people chatted for hours, discussing AI safety, model competition, and the future of Anthropic.
Ultimately, Osiri and Kutcher decided to invest in Anthropic through Sound Ventures, which they co-founded.
This fund has previously invested in OpenAI. Kutcher also specifically greeted Altman and explained that he was ready to invest in his competitor.
Since then, Sound Ventures has invested approximately US$435 million in Anthropic.
According to shareholding estimates disclosed by Bloomberg, the shares corresponding to this investment were worth approximately US$7 billion before the IPO.
This is not an ordinary cross-border investment in the entertainment industry.
It shows that Anthropic’s early capital support did not come only from traditional Silicon Valley venture capital. Music agents, actors, technology entrepreneurs, and traditional institutional investors are all participating in this AI race at different stages.
What’s even more interesting is that Kutcher’s investment was not in a movie, nor in an entertainment technology company, but in a basic model company that had no proven business model at the time.
The investment boundaries of AI have gone beyond the technology industry long before it became a popular topic.
When the model gradually becomes a new productivity infrastructure, capital also begins to gather in it from various industries.
10. Anthropic goes public, what should OpenAI do?
Anthropic's listing means that the capital market finally has a model company that can continue to observe, disclose operating data on a quarterly basis, and accept transaction pricing every day.
In the past, investors could only judge the value of companies like OpenAI and Anthropic based on funding rounds, private equity deals, and various valuation rumors. But today, ANT’s daily stock price, trading volume, financial performance and shareholder transactions will become the new reference for the industry.
This is especially noteworthy for OpenAI.
Five years ago, the Amodei brothers and several colleagues left OpenAI and founded Anthropic. Today, Anthropic is the first to ring the bell.
The competition between the two companies has also extended from model capabilities, customer competition and computing power reserves to the public capital market.
Whether Anthropic's stock price can stand firm, whether its revenue growth can be sustained, and whether its huge investment in computing power can be converted into profits will all affect investors' judgments of other model companies.
但公开市场与一级市场最大的区别在于,后者可以不断为未来融资,前者每天都在检验此前的预期。
AI公司终于有了一个公开的价格,但有价格不等于有答案。
Anthropic接下来要证明的,是今天的估值能够被未来的经营成果支撑。
写在最后:
Anthropic的IPO,至少同时完成了三件事。
它创造了史无前例的融资纪录,让一批早期员工和投资人获得了巨额账面财富,也为前沿大模型公司建立了一个公开市场的估值参照。
但它也把AI产业尚未解决的问题集中暴露出来。
算力需要持续投入,模型要不断迭代,资本也期待获得回报,社会还在追问技术发展的边界。
这几件事并不总是朝着同一个方向前进。
AI已经证明了自己能够创造巨大的商业价值。如今,轮到它证明自己能够创造与投入相匹配的资本回报。
这才是Anthropic上市留给整个AI行业的真正考验。
敬原创,有钛度,得赞赏
根据《网络安全法》实名制要求,请绑定手机号后发表评论
快报
11:48
刷新纪录,中国实现6G频段最高码速率无线通信验证
马来西亚总理安瓦尔抵达会见现场
11:47
习近平会见泰国总理阿努廷
启元机器人与邮电器材消费电子事业部签署战略合作
11:46
A股午评:深成指、创业板指跌均超2%,贵金属、白酒板块表现低迷
11:35
字节相关负责人辟谣谢欣与张楠离职传闻:该信息不准确
11:32
国内期货主力合约涨跌不一,钯、沪银跌近9%
11:28
Meta拟推出Muse Charm,定价为300美元
11:27
现货黄金失守3520美元/盎司,日内跌近3%
11:23
中国超六成可持续发展指标接近或实现2030年目标
11:14
AI漫剧概念盘中震荡回升,果麦文化涨超10%
11:13
宝洁计划到2030年削减最多30%制造和物流员工
11:04
国家能源集团在鄂尔多斯成立氢能科技公司,注册资本约3.5亿
11:00
国内金饰克价跌破1100元
10:55
韩国交易所启动SIDECAR机制,暂停KOSPI程序化卖盘
10:54
猪肉概念盘中震荡回升,金新农走出6天3板
10:51
22家私募现身15只首批主动ETF,合计持有5.24亿份
10:48
阿里2027财年Q2:闲鱼日活同比增23%,订单量增29%
10:47
泰国总理阿努廷抵达会见现场
10:46
五部门印发2026版海洋产业投融资《指导目录》
扫描下载App




