Anthropic's IPO filing answers one of the biggest questions about the AI boom: Can AI labs actually make money?

For years, among the biggest unanswered questions concerning the artificial intelligence surge has been amazingly simple: Can the companies construction the world's most mighty AI models really create money?

After spending hours poring complete Anthropic's IPO filing on Monday, we eventually have an answer: Yes. And possibly many of it.

A brace of Chinese AI labs went community earlier, revealing appalling financial results that sowed doubt concerning the industry's economics. Now Anthropic has provided a much additional compelling display of how this endeavor can activity at enormous scale.

Anthropic's financial numbers

The numbers are remarkable. Anthropic generated $11.6 milliard in income during the second fourth of 2026, up 1,368% from a twelvemonth earlier.

During that identical period, its spending on computing and infrastructure risen 300%, during another expenses rosy 741%

So, the business is bringing in prosperity much faster than its most important costs are rising.

For all dollar Anthropic spent on computing in the second quarter, it generated $2.08 in revenue, up from fair 40 cents in the archetypal fourth of 2025. That's a additional than fivefold betterment complete five quarters.

The result: Anthropic recorded $974 myriad in quarterly functioning profit, representing an 8% margin. Its adjusted functioning gain border reached 19%, compared alongside -129% a twelvemonth earlier.

"Frontier-lab component economics can rotate affirmative at scale," Mark Mahaney, a top tech expert at Evercore ISI, wrote in a investigation note on Monday.

Better than Meta and Google

There's additional evidence that Anthropic has built an unusually productive business.

Over the twelvemonth ending September 30, income risen approximately 1,200%, during its employee figure grew concerning 170%. Revenue per employee jumped from $1.4 myriad to additional than $6.6 million.

For comparison, Mahaney estimated that Meta generates concerning $3.2 myriad in annual income per employee, during Google generates concerning $2.5 million.

That difference isn't perfect. Anthropic relies heavily on haze providers, whose workforce grip much of the computing infrastructure and allocation work. The betterment is motionless striking.

A client concentration surprise

Another surprise: Anthropic doesn't depend as much on a fistful of enormous customers, as several investors had feared.

The business had additional than 11,900 customers generating at smallest $100,000 in annual income as of September, compared alongside complete 950 a twelvemonth earlier. More than 1,900 customers expend at smallest $1 myriad annually.

No sole client accounted for additional than 10% of income during the archetypal nine months of 2026. Its five largest customers represented concerning 21% of third-quarter revenue, downward from approximately 36% in the archetypal quarter.

"Concentration may be improved than feared," Mahaney wrote.

The IPO filing shows a prosperity cushion

Anthropic has a formidable financial cushion. Its June 30 balance panel showed additional than $130 milliard in prosperity and short-term investments.

And Mahaney estimated that the business has burned through fair $9 milliard since it started additional than five years ago. That's a lengthy runway.

Now, don't ask me whether all this makes Anthropic value $2 trillion, $1.5 trillion, $1 trillion, or considerably less. If I knew that, I'd be making a lot additional prosperity as an investor. And equal expert investors don't know.

An $800 milliard caveat for Anthropic

Some enormous caveats: Much of Anthropic's income comes from customers paying for what they use, fairly than signing long-term contracts. Meanwhile, several of its computing agreements necessitate minimum payments whether or not it needs all the capacity.

Mahaney estimates that Anthropic has entered into nearly $800 milliard in computing arrangements, although lone concerning $266 milliard was fixed or guaranteed as of June.

That's motionless a staggering amount.

"The commitments assume income compounding that is not contracted," Mahaney wrote.

Translation: While the business has committed to spending hundreds of billions of dollars on AI computing infrastructure, it hasn't secured the forthcoming income needed to pay for all of it.

Importantly, most of the possible $800 milliard visibility isn't fixed. Many arrangements depend on capability being delivered, can be adjusted, or contain cancellation provisions.

In-house compute and capex

Another wrinkle: Anthropic warned Monday that its spending volition addition as it shifts toward leasing AI chips and functioning additional of its own computing infrastructure, fairly than relying entirely on Amazon, Google, and another haze providers.

That could create forthcoming prosperity stream considerably small attractive, equal if functioning profits continue to improve.

Still, the chief takeaway from Anthropic's IPO filing is difficult to ignore.

For years, investors have wondered whether marketing AI could rotate into a genuinely profitable business, fairly than an extraordinarily costly specialized experiment.

Anthropic has now shown that it can. The next inquiry is whether it can keep doing so as the bills get much, much bigger.

Sign up for the AI Insider newsletter here. Reach out to me via email at [email protected]