OpenAI’s annualized revenue stood at about $50bn at the end of September, roughly $20bn below the figure recently circulated among investors. The difference has unsettled markets and revived questions about how quickly demand for artificial-intelligence services is translating into recognized revenue.
It reflected an attempt to recast OpenAI’s sales using a method closer to Anthropic’s, whose reported revenue includes the full value of some transactions completed through cloud providers such as Amazon Web Services and Google Cloud. OpenAI records only its share of certain partner sales, creating a substantial difference in the headline numbers even though both companies follow GAAP accounting.
Annualized revenue is calculated by projecting a recent sales pace across a full year. That makes it useful as a growth indicator but sensitive to the period selected, particularly for fast-changing businesses. It also says nothing by itself about costs, cash usage or profitability, leaving investors without the standardized quarterly disclosures required of public companies.
The disclosure came as OpenAI was discussing a potential funding round of at least $30bn at a valuation of about $1. 4tn. The company raised $122bn in March at an $852bn valuation, while a subsequent $30bn Anthropic financing round valued that rival at $965bn.
Investors reacted immediately to the revenue comparison. The moves reflected the sensitivity of chipmakers and cloud infrastructure companies to expectations that AI demand will support extraordinary levels of technology spending.
He has linked the decision to concerns about AI safety, while the company and its competitors face scrutiny over autonomous systems, researchers’ warnings and calls from politicians in both major US parties for new rules.
But the episode shows how heavily private AI companies’ valuations and the broader technology trade depend on estimates that are difficult to compare until firms publish audited public-company filings.