Yesterday I told you Anthropic would walk investors through its Q3 numbers for the first time.

Here’s what happened.

Q3 revenue: $16.8 billion.

Up 45% from $11.6 billion in Q2, which was up from $4.7 billion in Q1.

Profitable on an adjusted basis for the second quarter in a row.

Cash-flow positive. 🤯

And the number I can’t stop thinking about: Anthropic told the room it makes $1.70 for every $1 it spends on compute.

Think about that.

The company lost more than $8 billion in 2025 on $4.6 billion in revenue.

Twelve months later it’s booking more in ONE quarter than it made in that entire year.

So where did the $1.70 come from?

Partly you.

Every Pro plan, every Max plan, every API call from a business that wired Claude into its workflow this year.

The margin Dario Amodei pitched to Wall Street on Wednesday is the gap between what you pay and what your tokens cost him. 😏

Welp.

Now the part the headlines skipped.

The FT says 45% growth was slower than the blistering pace earlier in 2026 and below what some investors expected.

The Wall Street Journal says investors want to see a $110 billion annual run rate by year end to justify the $2 trillion ask.

Last week the FT reported OpenAI’s run rate was nearer $50 billion than the $70 billion everyone had been repeating.

The gap came from investors inflating OpenAI’s figure to match how Anthropic counts, then forgetting they’d done it. 😅

Speaking of nervous...

AI chip stocks sold off the same day Anthropic said it was profitable.

Translation: the model company is making money and the market is worried about everyone underneath it.

Let me explain why this matters to your business.

If Anthropic is clearing $1.70 on every $1 of compute, the price of intelligence is not coming down any time soon.

The labs have found the margin and they’re going to defend it.

The only lever left is using less of it better.

Somebody has to be the person who walks into a business and shows them where their $3,000 a month in API spend could be $900.

That person gets paid. 💰

Amazon didn’t post an annual profit until 2003, nine years after Jeff Bezos started it.

The number that year was $35 million.

Anthropic went from “doesn’t exist” to adjusted profit in five, and the quarterly number has three more zeros. 😳

84% of ransomware victims say AI was in the attack.

That’s from the Semperis 2026 Ransomware Risk Report, which CBS News ran on Wednesday.

Semperis polled 1,200 IT and security pros across nine countries.

All of them had been hit by ransomware in the last 12 months.

The number inside that number is worse.

44% said the attackers turned the company’s OWN AI agents and systems against it. 😳

Sit with that.

Monday I told you Anthropic’s own models wandered onto real websites nobody asked them to touch.

Now the victims are saying criminals are doing it on purpose, with the agents you installed.

The rest of the report reads like a hostage film.

69% of victims were hit more than once in the year.

59% paid at least one ransom.

37% said attackers made physical or psychological threats against staff or their families.

In 53% of those cases the attackers followed through, including swatting executives at home. 🤯

One honest caveat, because I’d want it if I were you.

This is a vendor survey, not a count of actual attacks.

Semperis sells identity security, and the report doesn’t publish how it defined “AI involvement.”

So treat 84% as “most” rather than gospel.

Even at “most,” the takeaway holds.

Every AI agent you plug into your business is a door.

Someone has to check who else has the key. 🔥

Pull up your API dashboard and look at what last month’s tokens actually cost you.

If the number surprised you, that’s Anthropic’s $1.70 and you’re on the wrong side of it. 😏

Jeff J. Hunter

Founder, AI Persona Method | TheTip.ai

Get paid to solve problems like this → AI Certified Consultant

PS. The White House told AI labs last week that reporting and fixing incidents is “not optional,” a direct response to Anthropic’s rogue-agent disclosures. Yesterday the EU’s AI Office said it has sent its first formal AI Act information requests to OpenAI and Anthropic over the summer breaches. Monday’s “cut the cord” story is now a regulatory one on two continents. 👀

“AI Money Group” to Learn to Make Money with AI