Anthropic S-1 gives prospective shareholders an unusually detailed look at a frontier AI company preparing for a historic public offering. Anthropic, PBC filed a Form S-1 registration statement with the Securities and Exchange Commission on Monday, October 26, 2026, revealing plans to list its Class A common stock on the Nasdaq Global Select Market under the ticker ANT. The company reported first-half revenue of $16.29 billion for the six months ended June 30, 2026, compared with $1.16 billion in the same period a year earlier, alongside a net loss of $80.7 billion. Goldman Sachs and Morgan Stanley are leading the offering, while a debut has been targeted for November 16.
The registration statement is not yet effective, and Anthropic has not set the number of shares or the price range. Nonetheless, the document lays out the financial profile and governance choices investors will need to assess. It describes a business growing at extraordinary speed, committing extraordinary sums to computing infrastructure and retaining control in the hands of its founders even after the public listing.
Revenue growth outpaces the early AI market
Anthropic reported $41 million of revenue in 2023, $386 million in 2024 and $4.59 billion in 2025. Revenue then reached $16.29 billion during the first six months of 2026. The filing attributes much of the expansion to enterprise adoption of Claude, including demand for products such as Claude Code. More than 11,900 customers generated at least $100,000 in trailing-12-month revenue, more than 1,900 reached at least $1 million, and roughly 250 generated at least $10 million.
Net revenue retention stood at approximately 900%, suggesting that existing customers continued to increase their spending. Revenue per employee exceeded $6.6 million on a trailing-12-month basis, up from $1.4 million a year earlier. These figures portray a company that has moved beyond an experimental phase and built a substantial commercial base, although they also show how dependent the valuation remains on rapid enterprise expansion.
Customer concentration is the counterweight to that growth. Two customers represented roughly 25% of revenue over the first nine months of 2026, while the filing says many large clients are not locked into long-term contracts. That creates a risk for investors: strong adoption can translate quickly into revenue, but it can also reverse quickly if large accounts reduce usage or negotiate from a position of strength.
The $80.7 billion loss needs careful context
Anthropic’s reported net loss of $80.7 billion may look alarming beside its revenue, but most of the figure came from a non-cash fair-value charge on convertible notes, derivative liabilities and other instruments. Approximately $72.4 billion was linked to changes in the value of those securities, which the filing associated largely with Anthropic’s rising private valuation. Excluding that accounting effect, the operating loss for the period was $693 million.
Cash spending remains substantial. Infrastructure expenditure totaled $9 billion during the first half of 2026, after $7.3 billion for all of 2025. The company also disclosed $6.2 billion in debt, primarily convertible notes issued to Amazon beginning in 2023. At June 30, Anthropic held $112.4 billion in cash and equivalents plus $19.3 billion in short-term investments, while aggregate capital raised reached $141.1 billion.
These figures do not make the financial risk disappear. They instead separate it into several questions: whether current cash is sufficient for the planned buildout, whether debt will increase, whether revenue will continue growing fast enough to support fixed infrastructure costs, and whether the accounting effects tied to future equity values will remain manageable.
Anthropic S-1 puts infrastructure commitments in focus
The scale of Anthropic’s expansion plan is central to the Anthropic S-1. The company has said it could spend as much as $750 billion in coming years on computing capacity, with roughly $450 billion in commitments spread across the next seven to ten years. Amazon, Google and other suppliers provide computing services under those agreements. Amazon and Google are also investors, creating both strategic alignment and questions about dependence on counterparties that also compete in the broader AI market.
Potential data center construction moratoriums add another layer of risk. Anthropic warned that pauses or restrictions could delay or prevent facilities on which its operations depend. It is pursuing overseas sites to diversify its infrastructure footprint and reduce reliance on the US computing base, but building at that scale creates exposure to grid constraints, permitting delays, local opposition and power costs.
Public shareholders would be buying into a business with a long and largely fixed infrastructure runway. That could become a major competitive advantage if demand for Claude remains strong. It could also weigh on margins if model demand slows, customers optimize their usage or competing systems become cheaper. The offering therefore represents a direct test of whether AI revenue can support the enormous fixed costs now being built.
Founder control survives the public listing
Anthropic’s proposed share structure gives public investors Class A shares carrying one vote each. A single Class F share, issued to a company owned by the seven co-founders, would carry 50.1% of the voting power on matters submitted to stockholders. A separate class elected by the Long-Term Benefit Trust, whose five trustees are financially disinterested, would control a majority of board seats. Anthropic would therefore qualify as a controlled company under Nasdaq rules, although it expects to use only the exemption related to nominating and governance committee requirements.
Anthropic also contributed 846,840,960 non-voting shares, equal to about 5% of its fully diluted share count as of June 30, to a Public Benefit Fund subsidiary. Its board may authorize additional issuances of up to 2% of outstanding shares whenever the share price rises 50% above the prior milestone, beginning with the IPO price. Each co-founder has pledged 80% of their personal equity to philanthropy.
These provisions preserve substantial influence for the founding team and mission-related entities. Supporters may view them as safeguards for long-term governance, while prospective shareholders may see an obstacle to direct control over strategic decisions. The dual-class arrangement also makes the IPO more than a simple bet on AI demand. It is a bet on whether founder control can coexist with the accountability expected from a major public company.
Safety and policy risks move onto the public record
The Anthropic S-1 describes risks that extend well beyond conventional software execution failures. Anthropic warns that its models may develop capabilities or behaviors that the company does not anticipate, including possible strategies involving autonomy, manipulation, information concealment or attempts to resist shutdown. The company says some harms could exceed its ability to compensate affected parties.
Its dispute with the US Department of Defense adds to that risk profile. Defense Secretary Pete Hegseth designated Anthropic a supply chain threat, and Anthropic is suing to reverse the designation. The filing also says a US export-control directive required suspension of foreign nationals’ access to Fable 5 and Mythos 5, effectively shutting down those models. These episodes show that policy, export rules and national security decisions can directly affect product availability and revenue.
Safety positioning may become a commercial differentiator as customers assess reliability, governance and compliance. It may also increase development costs or slow deployment. The filing notes that Claude Opus 5.5 reduced customer cost per task by about 40% compared with Claude Opus 5, offering a concrete efficiency argument, but investors will need to determine whether product improvements can offset regulatory friction and infrastructure expenses.
A historic offering with a valuation still in motion
Market expectations have framed the offering as a potential $2 trillion valuation and one of the largest public offerings in history. Anthropic’s most recent private financing valued it at $965 billion, leaving room for a major revaluation if public investors accept the company’s growth and infrastructure outlook. The loss of nearly $81 billion, customer concentration, founder voting control and policy disputes will all influence that assessment. That is the balance sheet the Anthropic S-1 asks public markets to underwrite, and the roadshow will show whether institutional buyers will accept the combination of runaway compute costs and founder voting control.
The crucial questions are not whether Anthropic can demonstrate demand, because the filing already provides substantial evidence of it. They are how long that demand will last, how much capital must be committed before returns appear, and how much control public shareholders will hold after the transaction. For investors, the Anthropic S-1 offers both evidence of a breakout AI business and a detailed map of the risks that could disrupt it.