The Filing That Finally Shows a Profit
Anthropic made its S-1 public on Monday, October 26, and in doing so became the first standalone frontier AI lab to walk toward a public listing. The company intends to trade on Nasdaq under the ticker ANT. The Wall Street Journal reports a debut targeted for the week of November 16, with a raise of up to $100 billion and a valuation somewhere between $1.5 trillion and $2 trillion.
The headline most outlets grabbed was the loss. The more interesting line is buried three pages deeper: Anthropic made its first operating profit in the June quarter, close to $1 billion. For an industry that has spent three years being told it cannot convert enthusiasm into margin, that is the number that matters.
Revenue for the six months ended June 30 came in at $16.3 billion. In the same period of 2025, the company booked $1.16 billion. That is a 13-fold jump in twelve months, not a projection, not an annualized run rate stretched by a friendly chart, but reported revenue in a registration statement filed with the SEC under penalty of law.
The second quarter alone tells the story faster. Anthropic recorded $11.6 billion in sales between April and June, against $787 million a year earlier. The company's entire 2025 revenue was $4.6 billion. One quarter in 2026 beat the whole of the prior year by two and a half times, and it did so while turning a profit.
Operating performance improved on both ends of the ledger. The first-half operating loss narrowed to $693 million from $2.14 billion a year earlier. Revenue multiplied, and the operating gap shrank by two thirds. That combination is the thing every AI skeptic said could not happen on this timeline.
Where the customers actually are. The growth is not consumer curiosity. Nearly 12,000 organizations now spend at least $100,000 a year with Anthropic, up from fewer than 1,000 a year ago. That is a twelve-fold expansion of the serious-enterprise base in a single year, and it is the strongest available evidence that AI budgets have moved out of experimentation and into line items that finance teams defend.
Most of the money comes from businesses using Claude to write software and to grind through office work. Both are workloads with measurable labor attached, which is why procurement signs off. Coding assistants shorten review cycles. Document agents remove headcount from back-office queues. Nobody has to believe in a singularity to approve that invoice.
Reading the $80.7 Billion Loss Correctly
The net loss for the first half was $80.7 billion, which sounds like a company incinerating cash at a rate that should terrify any investor. It is not that. Roughly $72.4 billion of that figure comes from changes in the estimated fair value of convertible notes and other financing instruments, according to the filing. When private valuations climb, the embedded conversion obligations get marked up, and accounting rules force the increase through the income statement as a non-cash charge.
In plain terms, a chunk of that loss exists because Anthropic became worth more. It is the accounting mirror image of success, and it will look strange for exactly as long as the valuation keeps rising.
The filing does contain one warning that investors should not wave away. Anthropic told prospective shareholders that growth is likely to slow, possibly sharply. Third-quarter figures will land before the listing, and the roadshow will be built around them.
The Compute Bet Behind the Numbers
Behind the improving P&L sits an obligation stack that dwarfs it. Anthropic has committed hundreds of billions to computing capacity:
- Up to $120.5 billion with SpaceX for capacity tied to xAI infrastructure.
- About $110 billion with Amazon, an existing investor and cloud partner.
- An estimated $125 billion to $150 billion with Google, the other major backer.
Those commitments total somewhere between $355 billion and $380 billion. The filing states plainly that if demand falls short of projections, many of these payments must be made anyway, regardless of how much of the capacity Anthropic uses. That is the real risk in this document, not the paper loss.
Governance hands control to the founders. The seven co-founders, led by chief executive Dario Amodei and president Daniela Amodei, will hold just over half of shareholder votes after listing, so public investors cannot outvote them. Anthropic remains a public benefit corporation, obligated to weigh safety against returns, and each co-founder has pledged 80 percent of their shares to charity, a pledge the filing notes is not legally binding.
The prospectus runs roughly 400 pages, with more than 90 dedicated to risk factors. It warns that advanced systems could pose catastrophic or existential risks if deployed irresponsibly, and it flags political exposure after the Defense Department labeled the company a national security risk and barred it as a supplier.
Set that against the quarter that just happened. A frontier lab posted operating profit, thirteen times the revenue of a year ago, with twelve thousand enterprises paying six figures a year. Whatever the stock does in November, the argument that this business cannot make money is now a historical artifact, and public markets get to price the proof directly.
Comments