OpenAI’s Revenue Nears $50B, Below Earlier $70B Reports

OpenAI’s revenue story has acquired an important correction. The company told investors its annualised revenue was approaching $50 billion at the end of September, according to the Financial Times. That is about $20 billion below the $70 billion figure that circulated in reports only days earlier. For a company asking investors to back an enormous expansion of AI computing, the difference is too large to wave away.

It is also easy to misunderstand. This is not evidence that OpenAI suddenly lost $20 billion in sales. Axios reports that the higher number reflected an attempt to count certain sales through cloud partners on a gross basis, making OpenAI’s business easier to compare with rival Anthropic. The roughly $50 billion figure counts OpenAI’s own share of some partner sales instead. The two numbers describe different ways of looking at a run rate, not a before-and-after collapse in revenue.

Annualised revenue is itself a projection. A business takes revenue generated over a recent period and imagines that pace continuing for a year. It is useful for seeing momentum, especially at a fast-growing private company, but it is not the same as revenue actually earned over twelve months. Nor does it tell readers how much cash remains after the cost of training models, serving users and securing data-centre capacity.

The distinction matters because the $70 billion number had begun to shape the conversation about OpenAI’s valuation and its ability to finance future growth. Our earlier look at its reported funding talks cited that widely reported figure while noting that outsiders could not inspect a private company’s books. The newer reporting gives that caveat fresh weight. Anyone repeating the larger number now should explain how it was calculated.

The issue is also a warning against treating every AI company’s headline revenue as directly comparable. A customer may pay through a cloud platform, leaving the model provider and its partner to recognise different portions of the transaction. Axios says both OpenAI and Anthropic’s approaches can comply with accounting rules. What investors need is a consistent view of the money each company retains, the costs attached to those sales and the definitions behind the figures in a fundraising pitch.

Markets reacted to the report on Thursday, with technology shares under pressure. That does not settle the longer argument about whether AI demand will justify spending on chips and data centres. A revenue run rate approaching $50 billion would still describe a huge commercial business. But the latest discrepancy shows why a big number, however exciting, cannot do the work of a clear financial statement.

For OpenAI, the immediate challenge is one of clarity rather than a proven loss of customers. Investors and the wider industry will want to know which sales are being counted, how much flows to partners and what remains after the substantial cost of delivering AI at scale. Until those details are public, even a spectacular revenue headline needs a careful footnote.