Anthropic’s extraordinary growth is beginning to answer one of the biggest questions hanging over the artificial intelligence boom: can the companies building expensive AI models actually make money?
The Claude developer reportedly generated more than $11.5 billion in revenue during the second quarter of 2026, up more than 14-fold from the $787 million recorded in the same quarter last year. Revenue also more than doubled from $4.73 billion in the first quarter.
But $11.5 billion in sales does not mean $11.5 billion in profit
The important figure is considerably smaller. Anthropic had previously told investors it expected around $559 million of adjusted operating profit for the quarter.
That would represent a margin of roughly 5% on $11.5 billion of revenue. Reuters reported that this measure includes the cost of training new models but excludes stock-based compensation.
No detail yet
There is an important qualification, however. Anthropic is a private company and does not yet publish the detailed audited accounts that investors would normally use to establish net profit.
The latest reports therefore point to positive adjusted operating income, rather than confirming $559 million of conventional net profit.
AI is expensive
That distinction matters because running frontier AI models remains extraordinarily expensive. Computing power, data centres, chips, model development and staff can consume vast sums.
Nevertheless, the shift is significant. Anthropic appears to be moving from an AI company dependent on enormous amounts of investment capital towards one capable of generating operating profits from its own customers.
So, what was the profit?
- Revenue: $11.5bn+
- Adjusted operating profit: approximately $559m
- Adjusted operating margin: roughly 4.9%
- Actual net profit: not publicly disclosed
- GAAP profit: we cannot say that Anthropic made $559m of conventional net profit
For investors contemplating a potentially huge IPO, that may be just as important as the spectacular revenue growth.
