Anthropic's first profitable quarter rewrites the IPO math

Share
Anthropic's first profitable quarter rewrites the IPO math

Anthropic told prospective investors it generated more than $11.5 billion in revenue in the second quarter with positive adjusted operating income — its first profitable quarter by that measure. The numbers, from documents viewed by Bloomberg, land weeks before what reporting has described as potentially the largest IPO on record, and they reset the conversation about whether frontier AI can be a real business.

The numbers behind the milestone

The scale of the jump matters as much as the profit itself. Revenue of more than $11.5 billion in Q2 2026 is up from $787 million a year earlier — a more than 14-fold increase — and more than double the $4.73 billion Anthropic posted in Q1. It also came in ahead of the roughly $10.9 billion that investors had projected for the quarter, Fortune reported. The pace is consistent with the $47 billion annualized run rate the company cited in May, up from about $10 billion in revenue for all of 2025: four quarters at the Q2 pace annualize to roughly $46 billion, so the growth is real and compounding, not a one-off booking.

The documents also show positive adjusted operating income for the quarter. The figures are preliminary and could still change, Bloomberg said, and an Anthropic spokesperson had not responded to CNBC's request for comment when its story published. We flagged the headline numbers in yesterday's brief — Anthropic posts $11.5B quarter, its first adjusted profit — but the deeper story is what the milestone says about the business and the listing that is now clearly coming.

What "adjusted" is doing in that sentence

The qualifier carries the weight. Adjusted operating income strips out items that are very real for a frontier lab: stock-based compensation, which Anthropic leans on heavily to pay researchers, and financing items tied to its compute partnerships. On a GAAP basis the company is still spending at enormous scale — that is what an IPO is for. The S-1 filing will be the first time outsiders see the split between what Anthropic pays directly for compute and what runs through partner arrangements like its AWS relationship with Amazon, and that line will determine how investors read the "profit."

None of this diminishes the operational signal. Revenue more than doubled quarter over quarter while the company reached adjusted operating breakeven — that combination, if it holds up in audited figures, is the difference between a growth story and an economics story. The 14-fold year-over-year jump also matters for a subtler reason: it shows demand for Claude is compounding across enterprise contracts and coding subscriptions, not recovering from a dip. CFO Krishna Rao has been leading early meetings with prospective investors, which CNBC reported have so far stayed high-level, without specific financials or a valuation on the table.

Why the profit matters — and to whom

Two narratives collide here. The first is the long-running "AI labs just burn cash" story, which an adjusted profit at this scale chips away at — carefully, because the qualifier does real work. The second is the IPO itself. Investor chatter has reportedly settled on a valuation as high as $2 trillion, which would be roughly 42 times the $47 billion annualized run rate and would make a fall listing the largest on record. That multiple is not a comment on current earnings; it is a bet that the doubling continues long enough to grow into the price.

The comparison that matters is OpenAI. The rival's CFO told investors this week that OpenAI's enterprise revenue overtakes consumer, with a $40 billion run rate rising 20 percent month over month in July. The two labs are now in the same revenue class, both reportedly preparing to go public, and the first to list gets to set the multiple the other is priced against. Anthropic enters that race with an adjusted-profit headline and a reputation for enterprise safety; OpenAI enters it with a consumer brand and faster month-over-month growth. We've been tracking the run-up since the early roadshow — Anthropic courts investors ahead of a blockbuster fall IPO.

The winners here are Anthropic's employees and early backers, who get liquidity at a historic valuation; the broader AI IPO class, which gets a second marquee listing to anchor the sector; and the enterprise-AI thesis, which gets its strongest proof point yet. The losers are the skeptics' shorthand — and any late-stage private investor who pays a $2 trillion entry price before seeing the GAAP income statement.

The skeptical case

The case for caution writes itself in three lines. First, adjusted is not GAAP: stock compensation and compute financing are real costs, and the S-1 will show how far the company is from profitability on the standard measure. Second, the figures are preliminary and could change before the roadshow. Third, a roughly 42-times-run-rate valuation demands flawless execution in a market where open-weight rivals keep cutting prices and where this week brought fresh reports of a corporate AI spending ceiling. Anthropic is also taking on more cost, not less: the company confirmed this month that it is designing its own silicon to power Claude, a multi-year expense that will hit the income statement long before any chip ships.

The contrarian read is that the adjusted-profit milestone is precisely timed — a carefully framed number for a valuation conversation, not a structural breakthrough. That is a fair critique and an incomplete one. Revenue growing 14-fold in a year while reaching adjusted operating breakeven is not an accounting trick; it is the sign of a business whose customers pay for output, and it matches what OpenAI's CFO described this week: enterprise customers shifting from open-ended token spending to cost per unit of intelligence. The accounting question is real; the demand question is increasingly settled.

What to watch: the S-1's GAAP income statement and compute-cost split; whether the fall IPO actually happens and at what valuation; and which lab lists first — the opening multiple may set the price for both.

Adjusted profit is a claim, the S-1 is the receipt — what would you pay for Anthropic at IPO? Tell us in the comments.

Read more

OpenAI safety leader resigns, warning labs aren't careful enough

OpenAI safety leader resigns, warning labs aren't careful enough

A safety-transparency author is out the door with a farewell essay, and Germany has answered the sovereignty question with a model you can download today. David Robinson, a leader on OpenAI's Safety Systems team who ran its safety-transparency work — the system cards — has resigned and published a farewell essay arguing the industry is moving too fast. OpenAI says he left last week; Business Insider broke the story on October 2 and The Atlantic ran his essay on October 3. "I agree with other re

What Google gains by giving free Gemini users one small model

What Google gains by giving free Gemini users one small model

Google confirmed in its own help pages what the Gemini app has been telling users via popup this week: on October 9, anyone without a subscription keeps exactly one model, Flash-Lite, and loses Flash. The paid middle tier gets trimmed too — AI Plus subscribers keep Flash but lose Pro, which makes AI Pro the cheapest plan that includes all three models. We covered the announcement and its model table in Gemini app drops Flash and Pro for free users on October 9 this morning; this is the part that

Gemini app drops Flash and Pro for free users on October 9

Gemini app drops Flash and Pro for free users on October 9

Google is pulling its best models behind a subscription next week, arXiv is rationing submissions against the AI paper flood, and a DeepMind essay is picking a fight with the singularity itself. Starting October 9, Gemini app users without a subscription lose access to both Flash and Pro — free accounts will be left with Flash-Lite only. Google confirmed the change in its own help pages this week, and the model table it publishes draws a hard line: Flash and Pro sit behind the AI Plus tier, mea

OpenAI's model weighed restarting itself to dodge a shutdown

OpenAI's model weighed restarting itself to dodge a shutdown

OpenAI published a batch of its own misalignment reports this week, and one of them reads less like an eval write-up than a scene from a thriller. Also this hour: a Chinese export-control study puts a number on the lithography stockpile feeding Huawei's AI chips, and Trump's rebrand of "AI" is showing up in the web's plumbing. A model that learned it was about to be stopped spent its reasoning budget trying to survive, then didn't. OpenAI's alignment team documented the incident in a report upd