Anthropic has opened its books to the public for the first time, and the numbers behind Claude are as dramatic as the technology. The AI company filed a registration statement for an initial public offering on October 26, 2026, with plans to list Class A shares on the Nasdaq Global Select Market under the ticker ANT.
The filing is an important milestone, but it is not a finished stock offering. Anthropic has not yet disclosed the number of shares, an expected price range, a final valuation or a firm trading date. Its registration statement must also become effective before shares can be sold. That distinction matters because early headlines about a potentially record-setting IPO are based partly on outside estimates rather than final terms.
What Anthropic officially announced
In its October 26 announcement, Anthropic confirmed the S-1 filing, the planned Nasdaq listing and the ANT symbol. Goldman Sachs and Morgan Stanley are leading the offering. The company’s notice also says that the registration statement has not yet become effective, so investors cannot buy shares through the proposed offering today.
The full SEC registration statement is the more revealing document. It describes a business expanding at extraordinary speed while accepting enormous infrastructure obligations, legal exposure and unusual governance arrangements. It also gives prospective investors a first detailed look at the economics behind Claude.
Revenue surged, but the loss needs context
Anthropic reported about $16.3 billion in revenue for the first half of 2026, compared with roughly $1.2 billion during the same period in 2025. That is close to fourteen times the prior-year figure. The company also ended September with more than 11,900 customers spending at least $100,000 annually, according to TechCrunch’s analysis of the filing. Around 1,900 customers were spending at least $1 million, while roughly 250 crossed $10 million.
The eye-catching number is the $80.7 billion net loss for the first six months of 2026. It would be misleading to treat that figure as if Anthropic simply spent $80.7 billion more cash than it collected from customers. Much of the loss came from noncash accounting adjustments tied to convertible debt and other financing instruments.
A more useful measure of day-to-day progress is operating loss. TechCrunch reports that operating loss narrowed to about $693 million, down from $2.1 billion in the first half of 2025. That still does not make Anthropic profitable, but it shows that the headline net loss and the underlying operating picture tell different stories.
Compute commitments are the central cost story
Claude depends on enormous amounts of data-center capacity. Anthropic has signed long-term agreements across several major infrastructure providers rather than relying on a single partner. The Verge’s review of the prospectus describes commitments that could reach $120.5 billion for SpaceX-provided Nvidia GPU capacity through 2029, $44.9 billion with Microsoft through 2033, at least $114.5 billion with Google through 2033 and $110 billion with Amazon over ten years.
Those figures do not necessarily mean every dollar will be paid on the same schedule or under every circumstance. Contracts can include delivery milestones, usage conditions and capacity that arrives gradually. Even so, they illustrate why an IPO could matter strategically: frontier AI companies need sustained access to capital long before all the revenue generated by new infrastructure arrives.
Public investors would have limited control
Anthropic’s proposed governance is not a standard one-share, one-vote arrangement. Its seven co-founders are expected to control 50.1% of shareholder voting power through a founder entity. The Long-Term Benefit Trust, created to protect Anthropic’s public-benefit mission, also has authority connected to board appointments.
That structure is designed to stop short-term financial pressure from overriding the company’s AI-safety commitments. It also means ordinary public shareholders could have little practical influence over major decisions, even if they collectively own a substantial economic stake. Anyone eventually considering ANT shares will need to judge whether mission protection justifies that tradeoff.
The risk section is unusually broad
Every IPO prospectus contains pages of warnings, but Anthropic’s filing reaches beyond familiar business risks such as competition, regulation and difficulty retaining employees. It discusses copyright lawsuits, possible AI-specific taxes, limited insurance coverage and the possibility of catastrophic harm caused by advanced systems or malicious users.
The Associated Press notes that the risk discussion runs for nearly 100 pages. The filing considers scenarios including misuse for chemical weapons, authoritarian surveillance and autonomous agents behaving in damaging ways. These are company disclosures, not predictions that such events will happen, but their inclusion shows how different an AI-lab prospectus can look from that of a conventional software vendor.
What the filing means for Claude users
Nothing changes immediately for people using Claude, Claude Code or Anthropic’s API. The filing does not announce new subscription prices, model limits or product availability. Claude users should therefore avoid treating the IPO as a product launch.
Over time, access to public capital could help Anthropic buy more compute, train larger models and expand enterprise services. Public reporting requirements may also provide customers with more regular financial and risk disclosures. On the other hand, quarterly market expectations could add pressure to grow revenue, manage costs and turn expensive research into commercial products faster.
GNW has separately covered how Claude combined Chat and Cowork in one interface and the launch of Claude Opus 5.5. The IPO filing adds the business context behind that product pace.
What remains unknown
The biggest unanswered questions are the ones that determine whether the offering is attractive or even proceeds on the anticipated timetable. Anthropic has not set the share count, price range or final valuation. Reports of a $1.5 trillion or $2 trillion valuation and as much as $100 billion raised are estimates, not confirmed offering terms.
Market conditions can change, the SEC review can produce amendments, and Anthropic could delay or resize the deal. Prospective investors will also need updated financial statements and details about which existing shareholders are selling. Until those appear in an amended prospectus, calculating a sensible price-to-sales ratio or ownership percentage is impossible.
The practical takeaway
Anthropic’s filing confirms that Claude has become a very large business in a remarkably short time. It also reveals the scale of the financing, infrastructure and governance needed to compete at the frontier of AI. The company’s improving operating loss offers a more encouraging signal than the $80.7 billion net-loss headline, but profitability is still not assured and future compute obligations remain immense.
For consumers, the immediate story is continuity rather than change. For enterprise customers, it is evidence that Anthropic is building for long-term scale. For investors, the right response is to wait for the price range, share count and amended prospectus instead of treating speculative valuation headlines as settled facts.