#直通IPO第三期Anthropic $69.3 Billion IPO: AI Bubble or the Next Decade’s Major Bull Run?

Anthropic’s IPO roadshow may be the most dramatic moment in the 2026 capital markets.
The day before, the market was discussing an offering plan seeking to raise about $69.3 billion at a valuation approaching $2 trillion; the next day, the filing was withdrawn and replaced with a more restrained offering range of $70–$78 per share, with the public offering’s fundraising cap adjusted to about $64.4 billion. Combined with a $26 billion private placement, the total could reach about $90.4 billion. Valuation expectations also fell from $2 trillion to $1.1 trillion–$1.4 trillion. This was not simply a pricing mistake, but a market repricing of AI assets.

Anthropic’s prospectus is half myth, half bill.
The myth is dazzling: as of the end of October 2026, annualized revenue was about $81 billion; third-quarter revenue was $16.96 billion, up more than 11-fold year over year; and the number of enterprise customers paying more than $100,000 annually reached 11,900. From $41 million in 2023 to $81 billion today, this growth curve has almost outpaced Google’s and Meta’s pre-IPO growth.
But the bill is just as staggering: the company has signed commitments to purchase about $900.9 billion in computing capacity and infrastructure, with about $70.7 billion due in the next 12 months. In the third quarter, the company’s operating profit turned into a $567 million loss from a $974 million profit, as it temporarily procured computing capacity at higher prices.
The structure of the participants is more complex still: seven giants—Google, Nvidia, AMD, Broadcom, Micron, Microsoft, and TSMC—have subscribed to a combined $26 billion in private placements. They are customers and shareholders of Anthropic, as well as suppliers of computing capacity and chips. This creates a closed loop: Anthropic trades equity for computing capacity, computing suppliers trade orders for equity, and companies across the industry chain boost one another.

Is this a bubble or a new cycle? More accurately, it may be both.
The bubble side is that valuations are still based on the assumption of “continued ultra-high growth in the future.”
If AI model prices keep falling, open-source models catch up, and corporate budgets tighten, trillion-dollar valuations will look burdensome. Especially as sequential revenue growth falls from 144% to 47%, the market has already begun recalculating its margin of safety.
The new-cycle side is that AI is no longer just a concept; it is entering enterprise workflows.
Claude Code, Claude API, enterprise knowledge bases, and automated workflows are moving from “trials” to “paid use.” When a company’s annualized revenue exceeds $80 billion, it can no longer be simply classified as a “story stock.”
The real question is not whether AI has value, but where that value will accrue.
Over the past two years, capital markets rewarded “more powerful models”; going forward, the market will care more about “who can make more money.” Can computing costs come down? Can customer retention be sustained? Can the application layer generate stable cash flow? These are the factors that will determine whether AI enters a major bull run or faces a valuation shakeout.

For ordinary investors, a more practical approach is to treat it as a barometer of the AI cycle: if Anthropic ultimately prices its shares conservatively and the overallotment option is exercised, that would indicate continued confidence in leading AI assets, and sectors such as computing capacity, chips, and cloud services could continue to benefit; if the offering price is lowered and subscriptions are tepid, it could mean the AI sector is entering a period of stress testing, making it more dangerous to chase concept stocks; if the stock fluctuates sharply after listing, ordinary investors should be even more careful to avoid emotional buying, especially bets made with leverage or concentrated in a single asset.
For most investors, rather than chasing the halo of the “largest IPO in history,” it is better to focus on AI applications that can genuinely generate cash flow, reduce costs, and improve efficiency.
The next wave of AI opportunities may not lie in the model companies with the highest valuations, but in the application layer that turns AI into a productivity tool.
Anthropic’s IPO is not just a pricing experiment for the capital markets, but also a watershed moment as the AI industry moves from “telling stories” to “reporting earnings.” Trillion-dollar valuations can be exciting, but what is truly worth watching is always cash flow, customer spending, and the path to profitability. The trend will not disappear, but the people most likely to lose money in it are often those who see only the valuation and not the bill.

👇What do you think? Welcome to share your thoughts in the comments～$ANTHROPIC