Anthropic Prices at $83 and Raises $81B, Setting a New IPO Record on Nasdaq
The Claude maker becomes the first frontier AI lab to go public, pricing above range

Anthropic priced its initial public offering at $83 per share on November 18, raising approximately $81.34 billion across 980 million shares — the largest public offering in recorded history — and began trading on the Nasdaq Global Select Market on Thursday under the ticker symbol ANT. The debut surpasses SpaceX's June 2026 record by several billion dollars and marks the first time a frontier AI laboratory has become a publicly accountable company. It also arrives at an unusually fraught moment: the day before pricing, a group of roughly 89 academics, including Nobel, Turing, and Fields Medal laureates, published a formal open letter — the MisAnthropic Statement — accusing Anthropic of pursuing recursive self-improvement while marketing itself as a safety-first organization.
The $83 offer price landed above the $70–$78 range outlined in Anthropic's revised prospectus, and the offering itself expanded from 825 million shares to 980 million — both signals of institutional demand that exceeded expectations after weeks of market turbulence. Of the 980 million shares, Anthropic sold 956.35 million itself, with the remaining 23.65 million sold by existing shareholders; proceeds from the shareholder-sold portion do not flow to the company. Goldman Sachs and Morgan Stanley led underwriting, and the offering is expected to close November 20, subject to customary conditions, according to Anthropic's S-1 registration statement. Underwriters also hold a 30-day option to purchase up to 147 million additional Class A shares — a greenshoe provision that could add several billion dollars more if the stock trades above the offer price in the coming weeks.
At the IPO price, Anthropic carries a fully diluted valuation of approximately $1.5 trillion. That figure fell short of the $2 trillion some investors had projected earlier this year but represents a dramatic upward revision from the $965 billion valuation established in Anthropic's May 2026 Series H round. It also eclipses the $1.25 trillion private valuation OpenAI secured in October 2026 — meaning the first frontier AI lab to go public has opened at a higher market capitalization than its primary rival, which has delayed its own IPO until at least 2027.
A Circular Capital Structure Where Investors Are Also Customers
Anthropic's shareholder structure after the offering and a concurrent $26 billion private placement tells a story that standard IPO narratives rarely surface. Amazon, which has invested approximately $18 billion in Anthropic through multiple tranches, holds roughly 14.7% of the company post-offering. Google, which has invested roughly $33 billion, holds approximately 13.7%. Together, they own close to 28.4% of Anthropic — but they hold Class N shares, meaning they carry economic rights with no voting power. The seven co-founders retain voting control through the company's tiered share structure, which insulates them from the governance pressure public shareholders would normally be able to exercise, as detailed in Anthropic's prospectus.
That ownership distribution matters far more than a typical IPO investor table because Amazon and Google are not passive financial backers — they are simultaneously Anthropic's largest cloud compute providers and its primary distribution channels. Amazon Web Services and Google Cloud together accounted for roughly 47% of Anthropic's 2025 revenue, up from 32% in 2024 and 11% in 2023. Anthropic paid approximately $350 million in distribution fees back to these same two companies last year. The relationship is structurally circular: Amazon and Google invest capital into Anthropic, Anthropic pays them for compute and customer distribution, and those customers generate the revenue that justifies the valuation that attracted the investment in the first place, as disclosed in Anthropic's S-1 filing.
Alongside the IPO, Anthropic completed a concurrent private placement of approximately $26 billion in Class N non-voting shares to a group that includes AMD, Broadcom, Micron, Microsoft, and NVIDIA. No single venture capital firm holds more than 5%; Sequoia Capital, former Google CEO Eric Schmidt, and Singapore's sovereign wealth fund GIC each hold roughly 1%. The breadth of the semiconductor and hyperscaler participation signals that Anthropic's infrastructure and enterprise relationships now span essentially the entire AI hardware ecosystem.
Revenue Velocity Without Comparable Precedent
The financial picture Anthropic disclosed in its S-1 is extraordinary by any technology benchmark, with caveats that require equally careful reading. Revenue grew from $386 million in 2024 to $4.59 billion in 2025 — a 12-fold increase. In the first half of 2026 alone, revenue reached $16.3 billion, and Q3 2026 added another $16.8 billion, implying an annualized run rate well above $60 billion by mid-year. Year-over-year revenue growth for the three months ending June 30, 2026 was 1,368 percent. Net Revenue Retention sits at approximately 900 percent — meaning the average customer is spending nine times what they spent a year earlier, a figure that far exceeds the typical enterprise SaaS benchmark of 120 percent, according to Anthropic's S-1.
Gross margin currently stands at approximately 40 percent, well below the 70–80 percent typical of mature software businesses, but the company has disclosed a target of approximately 77 percent by 2028. That gap reflects the enormous compute cost of running frontier AI inference at scale. Anthropic currently pays xAI — formerly Elon Musk's AI company, now part of SpaceX following a February 2026 acquisition — approximately $1.25 billion per month for access to the Colossus 1 GPU cluster, a facility housing roughly 220,000 NVIDIA GPUs, to handle inference and training workloads while its own infrastructure scales. Total compute spend for 2026 is approximately $19 billion, as disclosed in Anthropic's S-1 filing.
The loss figures are large in absolute terms but require context. Anthropic reported an $80.7 billion net loss for the first half of 2026, a number that is largely attributable to non-cash adjustments related to preferred stock conversions ahead of the IPO rather than cash burn. The company disclosed approximately $265.6 billion in future fixed financial commitments — predominantly long-term compute contracts — an obligation structure that reflects the capital intensity of maintaining frontier AI capabilities rather than a sign of near-term financial distress.
The Safety Contradiction That Public Markets Now Price
No other company in technology history has gone public while its own senior researchers have stated, on the record, that the technology under development may carry a greater-than-10 percent probability of catastrophic harm to humanity within a decade. Anthropic's S-1 devotes 80 of its 261 pages to risk factors. Among them: explicit acknowledgment of "catastrophic or existential risks to humanity" as a potential outcome of the company's own research program, as documented in Anthropic's S-1 risk factor disclosures.
The company's safety posture has visibly shifted as the IPO approached. In February 2026, Anthropic published Version 3.0 of its Responsible Scaling Policy, dropping its prior pledge to halt model training if adequate safety measures could not be guaranteed in advance — a commitment that had been central to the company's safety narrative since 2023. In July 2026, more than 1,100 employees across frontier AI labs — a letter Anthropic subsequently endorsed — called for the creation of independent tools to slow or pause AI development if necessary. Pretraining researcher Jacob Coxon resigned before the offering; Evan Hubinger, an Anthropic alignment researcher, stated publicly that he personally estimated a greater than 10 percent chance that AI could kill all humans within the next decade. NYC Comptroller Mark Levine wrote to Anthropic in October urging a pause to the IPO specifically because of these unresolved safety commitments.
The day before pricing, the MisAnthropic Statement — published with approximately 89 initial signatories including Nobel Prize, Turing Award, and Fields Medal recipients — argued that Anthropic was actively pursuing recursive self-improvement research while presenting itself to regulators and the public as a safety-focused counterweight to less careful AI developers. Recursive self-improvement, in which an AI system participates in designing and training its own successors without full human oversight at each step, represents the capability threshold that most AI safety researchers identify as the point at which controllability becomes structurally uncertain.
How the IPO Lands in an Already Turbulent Week
The offering priced into difficult conditions. Jensen Huang, Nvidia's CEO and arguably the most consequential single figure in AI hardware, resigned on November 12 amid reports of an extramarital affair with a woman U.S. intelligence agencies identified as a suspected Chinese intelligence asset. Nvidia is both a supplier to Anthropic's training ecosystem and a participant in the concurrent private placement; the board-level uncertainty at Anthropic's most critical hardware partner added volatility the IPO needed to absorb. That the offering priced above its marketed range anyway suggests that institutional investors resolved that uncertainty in Anthropic's favor, at least for now.
SpaceX's June 2026 debut — which raised approximately $86 billion — provided the most direct recent comparison. SpaceX shares declined from their post-IPO peak significantly over the following months, illustrating how quickly the market can reprice a highly anticipated public offering when early euphoria meets quarterly reporting realities.
What Quarterly Earnings Accountability Cannot Undo
Anthropic is now subject to a reporting cycle that its founders never faced as a private company. Every quarter, it will disclose revenue, margins, compute costs, customer concentration, and cash position to public shareholders who can sell their position on any trading day. The founders retain voting control through the share structure and cannot be removed by shareholder vote. But they can no longer avoid the public reckoning that quarterly earnings calls impose — not on financial metrics alone, but on the claims they have made about safety research pacing, recursive self-improvement timelines, and their use of the company's stated mission as a regulatory differentiator.
Whether that accountability proves meaningful depends on whether Anthropic's safety narrative can withstand the kind of disclosure that public markets require. The company enters public life at a 1,368 percent annual revenue growth rate, a fully diluted valuation that exceeds its nearest private rival, and with roughly $81 billion in fresh capital to accelerate the capabilities program its own researchers have called potentially catastrophic. That is the central wager of the ANT IPO — and it is now on the public record.



