On November 19, 2026, Anthropic will list on the Nasdaq Global Select Market under the ticker "ANT." The AI safety company, founded in 2021, is advancing what could become one of the largest IPOs in tech history.
According to the S-1/A amendment Anthropic filed on November 10, the public offering comprises 825 million Class A common shares at a price range of $70 to $78 per share, implying gross proceeds of up to about $64.3 billion and a fully diluted valuation of about $1.3 trillion to $1.45 trillion. Meanwhile, seven tech companies—AMD, Broadcom, Google, Micron, Microsoft, NVIDIA, and TSMC—have signed agreements to subscribe at the IPO price for a total of $26 billion in Class N common stock through a private placement, with the transaction scheduled to close immediately after the public offering is completed.
From Confidential Submission to Roadshow: Six Months of Pacing
Anthropic submitted a confidential draft S-1 to the U.S. Securities and Exchange Commission on June 1, 2026, publicly filed its registration statement on October 26, and then amended it twice on November 2 and November 10. The roadshow officially began on November 10, with the pricing date expected to be November 18. The entire timeline shows fairly precise pacing—nearly a month faster than the listing process of competitor OpenAI, which submitted a confidential S-1 on May 22 of the same year.
Market reaction shows order demand far exceeding supply by several times. According to TradingView data, index fund participation is estimated at up to 28% of issued shares, about $19 billion.
The Two Sides of the Revenue Numbers
The financial data Anthropic disclosed in its S-1 is the most closely watched core material of this IPO. Q3 2026 revenue was $17 billion, up more than 12x year over year; cumulative revenue for the first three quarters was $33.2 billion, a year-over-year increase of 1,206%. The number of enterprise customers with annual spending over $100,000 exceeded 11,900, also growing more than 12x year over year.
But another set of numbers is equally striking: the GAAP Q3 operating margin was -3%, and the full-year net loss was $41.97 billion. The company noted in its filing that the net loss includes a $21.9 billion non-cash charge related to convertible notes, derivative liabilities, and changes in SAFE valuations—in other words, the book loss is largely a byproduct of this round of valuation surge, rather than a direct manifestation of operating deterioration.
Excluding this non-cash item, Q3 adjusted operating margin was about 15%, still a considerable distance from the 35% long-term profit target presented in the company’s roadshow materials. Compute cost is the core pressure: Q3 compute and infrastructure spending accounted for about 58% of revenue, widening notably from Q2. The total committed infrastructure investment disclosed in the S-1 is about $895.4 billion, a figure that suggests a long capital consumption period remains between now and positive cash flow.
The Strategic Logic Behind the $26 Billion Private Placement
The private investment group of seven tech giants is the most noteworthy part of this IPO structure. These seven companies—NVIDIA, Google, Microsoft, TSMC, Broadcom, AMD, and Micron—play roles across the AI industry chain as compute suppliers, cloud platforms, strategic partners, and even potential competitors. Google was already a major Anthropic investor; it is adding $10 billion to buy 135.1 million Class N shares, with its total stake capped by agreement at 14.99%.
The design of the Class N shares is itself intriguing: these shares carry economic rights but no voting rights and are exempt from the standard six-month lock-up restriction. For buyers, this means locking in strategic stakes at the public market price without governance obligations; for Anthropic, it amounts to credit endorsement from core nodes of the industry chain before its formal listing, while maintaining the founding team’s absolute control over the company.
The implicit message of this structure is that Anthropic’s vulnerability in compute dependence and its strong market position are two sides of the same coin. NVIDIA is both Anthropic’s compute seller and will become its shareholder; this alignment of interests lowers uncertainty while also making it harder for external investors to independently assess the future evolution of the cost structure.
The Race with OpenAI: Anthropic Already Leads on Revenue
Anthropic completing its listing first is not only a lead in capital market pacing. According to multiple financial media reports, Anthropic’s Q2 2026 revenue was $11.5 billion, surpassing OpenAI’s $6.7 billion for the quarter for the first time and completing a key reversal in status in the history of the two companies. In Q3, Anthropic’s annualized revenue run rate was about $67 billion, while OpenAI disclosed annualized revenue of about $50 billion to investors.
The formation of this gap is directly related to structural demand in the enterprise market. Based on customer cases in the S-1, Eli Lilly rolled out Claude to 50,000 employees, Rakuten deployed more than 70 business-line agents, and legal tech company Legora used Claude Enterprise to expand company-wide usage from about 50 seats to 700 seats. Such large enterprise contracts are the direct driver of Anthropic’s leading revenue growth.
Valuation Compression: Why the Market Moved from $2 Trillion to $1.4 Trillion
In early October 2026 market expectations, institutional investors valued Anthropic at about $2 trillion. But the final S-1 price range implies a fully diluted valuation of about $1.3 trillion to $1.45 trillion—still a substantial premium to the $965 billion Series H valuation completed in May, but significantly contracted from earlier expectations.
According to multiple media reports, before the roadshow there was discussion of lowering the range further to $1.1 trillion to $1.3 trillion, which was ultimately not adopted. This process shows that institutional buyers did not unconditionally accept a high valuation underwritten by high AI growth, but actively weighed growth data against the compute cost structure.
Anthropic’s designation as a Public Benefit Corporation in its S-1 is not a decorative clause—this legal form requires the board to balance specific public benefit goals alongside profit, while giving founders a stronger defensive barrier in corporate governance. For capital seeking to maximize short-term returns, this structure itself is a discount factor; but for long-term funds that identify with the AI safety narrative, it provides a valuation framework distinct from OpenAI.
After the IPO, the Real Test Begins
Anthropic committed in its S-1 to a long-term adjusted operating margin target of 35%. To move from the current 15% to 35%, there are only two core levers: either continue improving compute efficiency—the S-1 discloses more than 4x efficiency gains compared with Q1 2025—or keep revenue growth above the pace of cost expansion.
The latter path is getting harder. Large-scale enterprise deployments mean higher operating costs and more complex SLA commitments; global differences in sovereign AI, regulatory compliance, and data sovereignty require Anthropic to maintain independent infrastructure nodes in every core market. At the same time, the $895.4 billion infrastructure commitment means capital expenditure will remain the main component of cash flow for years to come.
After listing, Anthropic’s quarterly revenue figures and compute cost ratios will be scrutinized quarter by quarter in the public market. The transparency brought by quarterly disclosure will force the company to continuously balance long-cycle AI research investment against short-cycle capital market expectations. This is a pressure structure Anthropic has never faced before as a private company.
© 2026 Winzheng.com 赢政天下 | 转载请注明来源并附原文链接