OpenAI has told investors that its revenue for 2026 is projected to reach $50 billion, based on sales through the end of September, according to The Guardian. That is roughly $20 billion below a $70 billion figure that circulated after information provided to investors the previous month. The difference has raised fresh questions about the pace of demand for artificial intelligence, but it does not appear to mean that $20 billion in booked sales suddenly disappeared.

The central issue is how revenue is counted. The Guardian reported that the higher figure emerged from efforts by OpenAI investors to compare the company more directly with Anthropic. Anthropic includes sales made through cloud partners such as Amazon Web Services and Google Cloud in its projected revenue, while OpenAI does not include equivalent partner sales in its own number. Put side by side without that qualification, the two forecasts can suggest a cleaner race than the underlying accounting supports.

Cloud transactions flow into two machines that count partner sales differently.
OpenAI and Anthropic reportedly treat cloud-partner revenue differently in their projections.

That distinction matters because private AI companies disclose less standardized financial information than public companies, while their revenue projections are treated as signals for the entire sector. Anthropic reached $65 billion in forecast revenue by the end of July, according to the report. OpenAI’s $50 billion projection may therefore look smaller, yet the companies’ different treatment of cloud-channel sales makes a direct comparison difficult.

Markets reacted sharply to the reported gap. The technology-heavy Nasdaq closed 1.4% lower on Thursday, while Nvidia fell 2.9%, Oracle dropped 5.5% and Micron declined 4.8%, The Guardian reported. Those moves do not prove that OpenAI’s revised figure caused every decline, but they show how sensitive investors have become to evidence that could challenge expectations of relentless AI growth.

A valuation bridge connects a revenue tower to a vast AI infrastructure landscape.
Huge private valuations make the definition of projected revenue especially consequential for investors.

The scrutiny arrives as OpenAI is reportedly in early-stage discussions to raise $30 billion at a valuation of about $1.4 trillion. The company’s most recent fundraising closed in March, when it raised $122 billion at an $852 billion valuation, according to The Guardian. At those scales, definitions of annualized or projected revenue are not a technical footnote: they influence the story investors tell about whether future sales can justify present valuations and enormous infrastructure commitments.

The competitive context is also shifting. OpenAI chief executive Sam Altman said last month that the company would not pursue a stock-market listing this year, citing AI safety concerns, while Anthropic is expected to move toward an initial public offering as soon as next month, The Guardian reported. These are prospective plans and may change. For now, they place extra weight on private disclosures and investor materials that outsiders cannot evaluate as fully as audited public filings.

The lesson from the $20 billion gap is less dramatic than a revenue collapse and more consequential than a simple correction. AI companies are growing quickly, but the metrics used to describe that growth are not always comparable. Until reporting conventions become clearer, headline forecasts should be read with their definitions attached. OpenAI’s $50 billion projection remains immense; the confusion around the $70 billion comparison shows why immense numbers still require careful accounting.