Anthropic files S-1 for Nasdaq IPO under ticker ANT, disclosing $16.3B half-year revenue

Anthropic moves toward a Nasdaq listing

Anthropic, the company behind the Claude API, filed a Form S-1 registration statement with the U.S. Securities and Exchange Commission on October 26, 2026, the first formal step toward an initial public offering. According to dev.to write-ups that examined the filing, the company intends to list Class A ordinary shares on the Nasdaq Global Select Market under the ticker ANT, with Goldman Sachs and Morgan Stanley named as lead banks for the prospectus.

Much is still undecided. The number of shares and the price range have not been determined, and no securities may be sold until the SEC declares the registration effective. No listing date has been announced either; dev.to cautions that mid-November timings circulating in the press are reporting, not a fact from the filing.

The numbers inside the filing

The S-1 gives the first detailed public view of Anthropic's finances and describes a business scaling unusually fast. As dev.to summarises, revenue reached $16.29 billion in the six months ended June 30, 2026, up from $1.16 billion a year earlier, with the second quarter alone contributing $11.56 billion. Consumption-based revenue, recognised as customers use the service and measured by tokens consumed, came to $13.1 billion, or 80% of the total.

The customer base widened dramatically over the same period: more than 11,900 customers spent at least $100,000 on a trailing-twelve-month basis as of September 30, 2026, against more than 950 a year before. Revenue is also not concentrated; no single customer exceeded 10% of revenue in the nine months ended September 30, 2026, and about 79% of third-quarter revenue came from customers outside the top five.

Profitability is murkier. Per dev.to, the company recorded an operating loss of $693 million for the first half of 2026 even though the second quarter alone posted $974 million in operating income, while the headline net loss of $80.75 billion for the half-year is dominated by $72.39 billion of non-cash fair-value changes on convertible notes, derivative liabilities and SAFEs.

One figure deserves caution: dev.to notes that press tallies of Anthropic's total compute agreements with Google, Amazon, Microsoft and others vary widely, because they are sums assembled by journalists rather than a number declared in the S-1.

Risk factors with developer relevance

The filing's risk factors are written broadly, but several map directly onto things Claude API users deal with. Compute capacity is the second item in the risk-factor summary: Anthropic warns that failing to secure adequate compute, or to steer it effectively between workloads, could leave it unable to train competitive models or serve customers properly, and that cloud partners can suffer outages, security incidents or capacity constraints. For developers, capacity problems surface as 529 overloaded errors and 429 rate-limit errors.

The S-1 also states that Anthropic has changed, and may again change, usage limits, rate limits and the capabilities of its models and products. dev.to reads this as confirmation that rate limits are a moving policy, and that client code should read them from response headers rather than hardcode them.

On model lifecycles, the filing concedes that customers who have embedded older models may resist migration, that accelerated or forced deprecation could drive churn, and that safety considerations may lead the company to limit or delay retiring older models. Current public policy promises at least 60 days of notice before a published model is withdrawn; dev.to notes that claude-sonnet-4-5-20250929 is scheduled for withdrawal on November 30, 2026, with claude-sonnet-5-5 as the recommended replacement.

Pricing may evolve too. The filing says competitive pressure could push Anthropic toward lower prices, shorter contract durations or alternative models such as outcome- or task-based pricing, and observes that its agentic offerings are structurally more compute-hungry than non-agentic ones.

What changes, and what doesn't

Nothing changes for API users on day one. dev.to emphasises that API keys, organisations and usage tiers are untouched, current prices and rate limits remain set by Anthropic's documentation and console rather than the prospectus, and the 60-day deprecation notice window still applies. The filing is a disclosure, not a product announcement.

Why it matters

The S-1 begins Anthropic's transition from a private lab into a company that will report on a quarterly cadence, giving developers and competitors a recurring view of pricing, compute spending and model strategy. The revenue mix matters most for people building on Claude: with 80% of revenue tied to token consumption, the API is the core business, which aligns Anthropic's incentives with developers' unit economics while making any future pricing experiment, such as outcome-based billing, something that lands directly in invoices. The risk factors also put on paper what practitioners already suspected: capacity, rate limits and deprecation are operational realities, and the filing is a useful prompt to build integrations that treat all three as variable.