OpenAI Tells Investors Annualized Revenue Is Near $50 Billion, Not $70 Billion

OpenAI has quietly reset expectations about its own size. According to financial documents shown to investors and reported by the Financial Times, the company disclosed that its annualized revenue was approaching $50 billion at the end of September — about $20 billion below the figure that circulated in media reports only weeks earlier.

The gap is not a sudden slowdown. People familiar with the matter say it stems from how the two leading AI labs count their money. Anthropic includes sales generated through cloud partners such as AWS and Google Cloud in its revenue line. OpenAI does not, because those arrangements work differently on its books. When investors tried to compare the two companies directly, they ended up comparing two different definitions of revenue. The lower number is the one OpenAI itself stands behind.

Even at $50 billion, the scale is remarkable for a company that was little known outside technology circles a few years ago. It also shows how quickly the AI industry has become a business of real budgets, real infrastructure and real investors, rather than a research project with a demo attached. Annualized revenue is simply the most recent month’s pace multiplied across a year, so it is a snapshot rather than a guarantee — a distinction that matters when growth is this fast.

The disclosure landed on a jittery day for markets. The Nasdaq Composite Index fell more than 1 percent, and the Philadelphia Semiconductor Index dropped about 3 percent, with Intel, Marvell, Micron and AMD all down sharply. Nvidia, the most valuable company in the sector and a key supplier of the chips that train and run AI models, slipped around 2 percent. When the market’s biggest buyer of computing power revises its own numbers, the companies selling it hardware feel the tremor first.

The timing is awkward for another reason. AI infrastructure spending is now one of the main engines of global demand for advanced chips, data center construction and electricity. Oracle, for instance, has been exploring unusual ways to power its data centers while gas pipelines are delayed, including trucking natural gas directly to server farms — a strategy it has already used near Salt Lake City and in Texas. Such measures underline how tightly the fortunes of software companies and energy suppliers have become intertwined.

For global readers, the lesson is less about one company’s bookkeeping than about how young this industry still is. Revenue definitions are not yet standardized, partnerships blur the line between a company’s own sales and its partners’ sales, and investors are learning to read the fine print in real time. A $20 billion difference between two descriptions of the same business is a reminder that in AI, the numbers are still being written.

OpenAI has not commented publicly on the documents. What is clear is that the company remains one of the largest private businesses in the world by revenue pace, and that its disclosures now move markets far beyond its own industry. The next few quarters will show whether the $50 billion figure was a pause in the narrative or a new baseline.