Key Highlights

  • OpenAI says annualized revenue was USD 50 billion at the end of September 2026, and that an earlier figure of about USD 68 billion included partners' gross revenue, according to press accounts of company statements.
  • OpenAI also expects annualized revenue of at least USD 70 billion by year end, a target that is a company expectation and not a result, and it is private, so it has no ticker.
  • Nvidia (NASDAQ:NVDA) is the public company most often named as a read-through, followed by AMD and Oracle, because these suppliers provide the chips and cloud capacity that OpenAI buys.
  • Chip stocks fell more than 4% in the week to October 9, 2026, and the semiconductor index dropped as much as 3.4% on the original report, so the clarification matters for sentiment.

What OpenAI Clarified

Reports on October 8 and 9, 2026 relayed company statements that OpenAI's annualized revenue was USD 50 billion at the end of September. The company said the earlier USD 68 billion figure, which some reports placed at USD 68 to 70 billion, counted partners' gross revenue, meaning sales before payments passed to others. The difference of about USD 18 billion, simple arithmetic on the two numbers, may therefore be a reporting basis and not a loss of customers. OpenAI also said it expects at least USD 70 billion by year end. Annualized revenue is a recent period scaled to a full year, so it is a run-rate and not audited income. These details come from press accounts and cannot be checked against a filing.

Why Nvidia Is Exposed

Nvidia (NASDAQ:NVDA) designs the accelerators used to train and run large AI models, and OpenAI is among the biggest buyers of that computing capacity, often through cloud partners. Investors worried that a smaller revenue base would limit how much OpenAI could keep spending. Nvidia is quoted near USD 229 in published market data. Advanced Micro Devices (NASDAQ:AMD) sells competing accelerators. Oracle rents cloud capacity to AI developers, so each carries a similar read-through. A clarification that revenue is not collapsing helps sentiment, but it does not prove that orders will keep growing at the pace implied by recent guidance. Customers can also change vendors over time, as many large buyers design their own chips.

Risks Still on the Table

The year-end target of at least USD 70 billion would require growth of about USD 20 billion from the September run-rate, another simple arithmetic point, and it is not guaranteed. Funding costs matter too, because Treasury yields touched about 5.35% intraday this week and AI infrastructure is often financed with debt. Customer concentration remains a risk for suppliers, since a few buyers account for much of demand. Watch supplier commentary from ASML on October 14 and Taiwan Semiconductor Manufacturing on October 15, and the Federal Reserve meeting on October 27 and 28. Holders of Nvidia (NASDAQ:NVDA) should read company filings before relying on any press account.

Where Things Stand

OpenAI has offered an explanation, and it may ease the worst fears, but it does not settle the bigger question of how fast AI customers can grow. Nvidia (NASDAQ:NVDA) benefits if the gross versus net explanation holds, since the revenue base looks less fragile than the first report implied. The figures are reported, not confirmed by a filing. The next real evidence will come from supplier results this week and from any OpenAI update on its year-end target.