Anthropic is still targeting a $2 trillion IPO valuation for November, and its leaked prospectus lists $4.6 billion of full-year 2025 revenue, $7.33 billion of compute and infrastructure spend in 2025, and a $42 billion net loss in 2025, according to Crunchbase News. Anthropic’s third-quarter 2026 revenue was reportedly $16.8 billion, up 45% quarter on quarter, implying an annualized revenue run rate of roughly $67 billion, according to Crunchbase News.
Prospectus figures: revenue, compute spend, net loss, and infrastructure commitments
Anthropic’s leaked prospectus states that the company generated $4.6 billion in full-year 2025 revenue, according to Crunchbase News. The same prospectus figures cited in the article state Anthropic spent $7.33 billion on compute and infrastructure in 2025 alone, which is more than its total revenue for that year, according to Crunchbase News.
The prospectus figures cited in the article also include a $42 billion net loss for Anthropic in 2025, according to Crunchbase News. The article says the roadshow will explain that $34 billion of the 2025 net loss was a noncash accounting charge tied to an adjustment to the estimated value of financing instruments that could eventually convert into shares, according to Crunchbase News.
The article states Anthropic expects to spend, at least, $518 billion over the course of a decade on AI infrastructure with six partners including Google, Microsoft and Amazon, based on a confidential IPO prospectus reviewed by Reuters, according to Crunchbase News. The article adds that about 80% of that $518 billion total is noncancelable or payable regardless of usage, according to Crunchbase News.
The article states Anthropic’s targeted $2 trillion IPO valuation represents 435 times its 2025 revenue, according to Crunchbase News.
Customer concentration and public safety concerns cited around the IPO
The article says one more figure in Anthropic’s S-1 is that a quarter of Anthropic’s 2025 revenue came from two unnamed customers, according to Crunchbase News. The article also describes that the New York City pension fund publicly asked Anthropic to delay its IPO over AI safety concerns, according to Crunchbase News.
The article identifies NYC Comptroller Mark Levine as the supervisor of New York City’s pension funds and says he sent a letter to Anthropic CEO Dario Amodei urging the company to pause the IPO entirely, according to Crunchbase News. The article says Levine cited risks that are also flagged in Anthropic’s prospectus, including increasingly autonomous AI systems that could behave in unexpected ways, create security problems, and, in the document’s language, pose “existential risks to humanity,” according to Crunchbase News.
The article says New York City’s pension funds reportedly hold roughly $7 million in Anthropic exposure today, and the letter raised that the funds could hold dramatically more if Anthropic enters the Russell 3000 post-IPO, according to Crunchbase News.
OpenAI IPO timing, chip selloff, and Gartner revenue versus investment figures
The article states OpenAI has said it will not go public in 2026, and it says CEO Sam Altman called an IPO “ill-advised” given AI safety concerns, according to Crunchbase News. The article describes OpenAI as the “second-most valuable AI company in the world,” according to Crunchbase News.
The article states that last week’s $2 trillion chip-stock selloff was driven partly by concern over AI-buildout costs, according to Crunchbase News. The article says Gartner projects that through 2030 there will be at most $3.6 trillion in revenue attributable to AI data centers versus $8.5 trillion needed to cover planned investment, financing costs and a 12% return, according to Crunchbase News.
As widely-known context, an S-1 is the U.S. Securities and Exchange Commission registration statement used in a public offering.
Pricing and margin claims tied to open-weight models and valuation tests
The article reports that enterprise adoption rose from 23% to 67% in 2026, with companies reporting 70%–90% cost savings, in a discussion of open-weight AI alternatives, according to Crunchbase News. The article states DeepSeek R1 offers frontier-level performance on most enterprise benchmarks at 7 cents per million input tokens, and it says independent analysis put Anthropic’s comparable input and output pricing at roughly 18 times and 23 times higher, respectively, according to Crunchbase News.
The article states DeepSeek’s weights are free to download, allowing enterprises to self-host without a model-provider API contract or per-token billing, and it says the capability gap with closed models reportedly closes within six months of a frontier release, according to Crunchbase News. The article says New Constructs called Anthropic’s proposed offering “an unprecedented test of investor gullibility,” and it calculated that justifying a $2 trillion valuation would require an immediate 10% after-tax operating margin sustained for a decade, extraordinary revenue growth, and $358 billion in annual after-tax operating profit by 2035, according to Crunchbase News.
The article also states that worldwide spending on AI is forecast to reach $2.59 trillion in 2026, up 47% year over year, according to Gartner, and that a July 2025 MIT NANDA report estimated that 95% of the organizations it examined were getting no return from generative AI, according to Crunchbase News.
