One Hundred Billion and Counting: Anthropic's Revenue Pace Dwarfs the Field as Its IPO Slips to November

There is a number that used to belong exclusively to the largest software companies on Earth, and as of this week it belongs to a startup that did not exist six years ago. The New York Times reported on September 18, 2026, that Anthropic is now pacing to generate more than $100 billion in annualized revenue this year — a run rate that has swollen by 50% in just two months, from the $65 billion figure the company disclosed in July, and that now stands at more than ten times where Anthropic exited 2025. Yahoo Finance, summarizing the Times report, framed the trajectory plainly: growth is “accelerating at almost unfathomable rates” as the company heads toward what could be the largest initial public offering in history.

The scale is difficult to contextualize precisely because there is almost nothing to compare it to. Anthropic was founded in 2021. By the end of this year, as Yahoo’s analysis noted, it could rank among America’s 50 largest companies by revenue and its ten largest by market capitalization — a climb that took incumbents like Microsoft and Oracle multiple decades. Sacra’s independent estimates track the same curve from below: roughly $9 billion in annualized revenue exiting 2025, $14 billion by February 2026, $19 billion in March, $30 billion in April, $44 billion in May, and $65 billion by July. The Times’ new figure extends the sequence past a milestone that, for a private company not backed by a state, has no real precedent.

What is doing the pulling

Two products account for the shape of the curve. The first is Claude Code, Anthropic’s agentic coding tool, which has become the default workhorse inside a striking share of professional software teams. The second is Cowork, the enterprise product that packages Claude as a general-purpose workplace agent. Together they have converted what was once a chatbot subscription business into something closer to infrastructure spending — line items that enterprises renew the way they renew cloud contracts, not the way they cancel app trials.

The engine behind both is the same: usage-based pricing attached to work actually performed. When Claude Code writes, reviews, and merges code, or when a Cowork agent completes a research or operations workflow, tokens flow and revenue follows. That structure explains why the revenue pace can compound this quickly without a corresponding sales-force expansion — the product’s own success inside existing accounts generates the growth. It also explains the concentration risk that accompanies the headline number: a business this young, growing this fast, is overwhelmingly weighted toward the single hottest workload in the economy (software development), and toward customers whose AI budgets are themselves still experimental.

The IPO slips — deliberately

The second half of the story is timing. The Wall Street Journal reported that Anthropic has pushed its IPO from October to November 2026, and the reasoning is not distress — it is arithmetic. Waiting a few extra weeks lets the company present third-quarter financials to public-market investors, results expected to show robust competitive positioning in the period following OpenAI’s September release of its GPT-6-class “Astra” model. When the roadshow begins, bankers and executives plan to gauge market sentiment while directly addressing the question every allocator is asking: whether the industry’s sudden turn toward slower, more cautious model releases will impair future growth.

The figures attached to the listing remain, by every prior standard, absurd. Public-market participants still anticipate a transaction valuing Anthropic at approximately $2 trillion and raising up to $100 billion — figures that would comfortably eclipse the record set by SpaceX’s debut. Early backers quoted in recent reports project the company could clear $110 billion in annualized revenue by year-end, and at a private gathering company leadership highlighted new initiatives such as the “Model Hardware Standard,” a specification letting software agents interface directly with physical devices like robotic arms.

The Amodei paradox

What makes this listing genuinely unusual is the posture of the person running it. CEO Dario Amodei spent September publicly warning that frontier AI risks demand a slower, more deliberate development pace — calls that were echoed, remarkably, by OpenAI’s Sam Altman and xAI’s Elon Musk. Anthropic is simultaneously: publishing safety warnings that argue its own product category deserves restraint; asking public investors for the largest valuation ever awarded; and assuring those investors that a moderated release cadence “will not materially impair revenue momentum,” as company advisers and shareholders contend.

Existing venture backers have largely reframed the safety posture as an asset rather than a liability. In their telling — reflected in the WSJ’s reporting — institutional investors view deliberate risk mitigation as prudent insulation against the governance and regulatory scrutiny that inevitably arrives once a company is public. There is a version of this argument that is sincere; there is also a version that is marketing. The truth is probably that being the AI lab that publicly called for slowing down is worth more in litigation and regulatory exposure avoided than it costs in narrative edge over rivals racing ahead.

The timing shift, sources told the Journal, was actually decided before the public slowdown debate erupted — a reminder that in banking, the calendar usually follows the financials, and the financials here gave Anthropic every reason to wait. A Q3 print continuing the trajectory from $65 billion toward $100 billion+ is the single most persuasive slide in what will otherwise be a very thick prospectus.

What to watch

The competitive backdrop is not standing still. OpenAI is reportedly targeting a $1.2 trillion valuation in early fundraising discussions while deferring its own public debut until 2027 — and its CEO has said his lab will not go public this year, citing the same safety concerns Anthropic is navigating. Some Anthropic investors privately caution that another OpenAI mega-round could absorb significant market liquidity, which makes the execution of the late-autumn listing genuinely critical: there is only so much capital willing to sit in pre-profit AI positions, and Anthropic wants it committed before anyone else claims it.

For everyone outside the cap table, the number to watch is not the valuation but the growth durability. A $100 billion revenue pace built on Claude Code and Cowork is extraordinary; it is also, by construction, a leveraged bet that enterprises will keep expanding agentic AI budgets through whatever macro and safety turbulence the next year brings. The November prospectus will be the first public, audited look at that bet — and at whether the fastest revenue ramp in corporate history has a floor under it.

Either way, the milestone stands on its own. Six years from incorporation to a $100 billion revenue pace is a record that may never be touched again — and the market will get its chance to price it within weeks.