Anthropic Files Confidential S-1 as Revenue Run Rate Hits $47 Billion

The race to ring the bell at America’s stock exchanges just got a second, equally formidable entrant. On June 1, Anthropic confidentially submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission, formally beginning the process of taking the Claude maker public. The filing came just four days after the company closed a $65 billion Series H funding round that pushed its private valuation to approximately $965 billion — surpassing OpenAI’s $852 billion private valuation and making Anthropic the most valuable startup on earth.

The timing is striking. OpenAI had submitted its own S-1 only days prior, setting up what Wall Street analysts are already calling the most consequential AI IPO showdown since Google and Yahoo jockeyed for internet dominance in the early 2000s. This time, the stakes are measured in the trillions.

Revenue That Defies Precedent

The headline numbers are genuinely unprecedented for a company at the pre-IPO stage. Anthropic’s annualized revenue run rate has crossed $47 billion — up from $9 billion at the end of 2025. That fivefold increase in less than six months represents one of the fastest revenue ramps in Silicon Valley history.

The engine driving that growth is Claude Code. Anthropic’s AI-powered coding assistant has become the dominant tool in the enterprise software development market, with adoption spreading across Fortune 500 companies in engineering, legal, finance, and research workflows. As organizations increasingly deploy Claude not just for code completion but for autonomous multi-step agentic tasks — running CI/CD pipelines, coordinating between internal APIs, and generating entire application scaffolds — consumption-based billing has dramatically inflated average deal sizes.

Agentic deployments on the Anthropic API, where Claude operates with tool use and long-horizon reasoning over enterprise data, now represent a growing share of total bookings. These workflows are stickier than simple chatbot contracts and command meaningfully higher price-per-token rates.

The $965 Billion Question

At its current private valuation, Anthropic is approaching the trillion-dollar threshold — a level that only four companies (Apple, Microsoft, NVIDIA, and Alphabet) have sustained in public markets. Whether public investors will accept that price tag depends heavily on one number that the confidential S-1 has not yet disclosed: net profit margins.

The company, like all frontier AI labs, operates with structural cost challenges. Training and serving the largest Claude models requires enormous compute infrastructure. In May 2026, Anthropic announced it would pay SpaceX $1.25 billion per month for access to compute at the Colossus 1 data center in Memphis — a facility built by Elon Musk’s xAI before its merger with SpaceX in February. That single line item amounts to $15 billion annually, suggesting that even at $47 billion in revenue, the path to profitability requires either further price increases or dramatic inference cost reductions.

Investors backing the Series H — a who’s-who of institutional capital including Altimeter Capital, Dragoneer, Greenoaks, Sequoia Capital, Capital Group, Coatue, and D1 Capital Partners — are betting that Anthropic can thread that needle before it trades on a public exchange.

Claude’s Expanding Product Surface

The S-1 process will force Anthropic to disclose, for the first time publicly, the breakdown between its API business, Claude.ai consumer subscriptions, and enterprise contracts. Industry observers estimate that the enterprise segment, which includes multi-year agreements with companies like Salesforce, McKinsey, and several major financial institutions, now accounts for the majority of revenue.

Notably, the S-1 was filed while Anthropic was still in limited preview of Mythos, an advanced successor model that has reportedly identified thousands of high-severity bugs requiring remediation before wider release. How Anthropic characterizes the safety and deployment timeline of its next-generation models in the public filing will be closely scrutinized by both investors and AI governance advocates.

Claude’s model family has also been expanding along the capability-cost frontier. The Claude 3.7 and 4.x families cover a range from ultra-fast, lightweight inference for real-time applications to deep-thinking reasoning modes for complex agentic workflows. This product breadth mirrors OpenAI’s strategy with the GPT family and positions Anthropic to capture spending at every tier of the market.

Joining the $3 Trillion IPO Race

Anthropic is now the third major tech player expected to list in 2026, alongside OpenAI and SpaceX. Together, the three companies represent a combined private valuation exceeding $3 trillion — an amount that would have been unthinkable even 18 months ago.

The sequence matters. SpaceX filed first, revising its S-1 to include the Google compute deal and a “friends and family” share carve-out as it prepares for a Nasdaq debut. OpenAI followed days later with its own filing. Anthropic’s rapid confidential submission signals it intends to compete for the same pool of institutional capital and does not want to be the third company asking investors to absorb a near-trillion-dollar valuation.

The confidential filing process gives Anthropic time to work through SEC comments privately — a privilege extended to “emerging growth companies” and issuers that have not yet filed a prior registration statement. A public S-1 with full financials, risk factors, and ownership disclosures will follow, likely in the early fall if the company maintains its current timeline.

What Investors Will Watch

When the public S-1 drops, several questions will dominate analyst coverage. First, how much of that $47 billion ARR is actual annual contract value versus trailing run rate extrapolated from a few exceptional months? Revenue constructed from a handful of massive contracts with unusual consumption spikes can be misleading. Second, how concentrated is Anthropic’s customer base? Heavy reliance on a few enterprise anchor customers creates churn risk. Third, what is the competitive moat as Google, Microsoft, and Meta all race to produce models that match Claude’s coding and reasoning benchmarks?

For now, the confidential S-1 filing is a signal: Anthropic believes the window is open, the revenue story is compelling enough to withstand public scrutiny, and it intends to be a public company before the current AI market enthusiasm cools. On current trajectory, the bell-ringing could come as soon as the fourth quarter of 2026.