2026-10-09

OpenAI's Annualized Revenue Corrected From '$70B' to '$50B' - Oracle Falls 5.5%, Intel 6%, Nvidia 3% as Nasdaq Drops for a Second Day

Written & reviewed by the TradeSmrt editorial team · 2026-10-09

In this article

What Happened
Why a Pure Accounting Correction Hit Stocks This Hard
What to Take Away From This
FAQ
Sources

What Happened

AI infrastructure stocks sold off hard on Thursday, October 8. Oracle (NYSE: ORCL) dropped 5.48%, Intel (Nasdaq: INTC) fell nearly 6%, Broadcom (Nasdaq: AVGO) slid 4.35%, and Nvidia (Nasdaq: NVDA) lost 2.94%. AMD fell in the same 4% range. The damage wasn't contained to individual names - it dragged down the broader tape. The Nasdaq Composite fell 345.35 points (1.3%, with some outlets citing 1.25%) to close at 27,193.34, its second straight down day. The S&P 500 slipped 36.41 points (0.5%) to 7,765.36. The Dow Jones Industrial Average, by contrast, actually rose 51.77 points (0.1%) to 51,231.64 - a sign this was a targeted selloff in AI and chip names, not a broad market rout.

The trigger wasn't an earnings report or a regulatory filing. It was a Financial Times scoop. The FT reported that, based on financial documents it reviewed, OpenAI's annualized revenue stood at roughly $50 billion as of the end of September. That's about $20 billion below the "nearing $70 billion" figure that had circulated widely just a month earlier, largely via Axios reporting in late September. CNBC subsequently confirmed the $50 billion figure independently, and once two separate outlets were saying the same thing, the market treated it less as a rumor and more as fact - triggering an immediate selloff across AI-linked names.

Here's the part that matters: the $20 billion gap wasn't a sign that OpenAI's business suddenly shrank. According to the reporting, the widely-cited $70 billion figure was never a number OpenAI itself disclosed - it was an estimate investors built themselves. The chain of logic went roughly: start from an estimate that OpenAI's monthly revenue hit the $4 billion range back in August, annualize it, then layer on OpenAI's own disclosure that revenue had grown 70% - and investors arrived at something close to $70 billion. The problem is that this calculation baked in an apples-to-oranges comparison with rival Anthropic. Anthropic's annualized revenue figure includes gross revenue booked through cloud partners like AWS and Google Cloud, while OpenAI's own figure explicitly excludes that partner revenue and reports net revenue only. Two companies, the same label ("annualized revenue"), two different accounting conventions - and the market had been comparing them as if they were apples to apples.

Why a Pure Accounting Correction Hit Stocks This Hard

On paper, this should have been a minor footnote: a recalculation, not a business setback. In fact, the FT's own reporting noted OpenAI's overall third-quarter revenue run rate grew 77% year over year, with its enterprise segment growing an even faster 107%. By any normal growth-company standard, that's a strong quarter. So why did the market react so violently to what amounts to a definitional cleanup?

First, the entire AI value chain sitting on top of OpenAI had already been priced using the unverified $70 billion figure. Oracle's stock has been re-rated over the past year largely on the strength of its massive cloud computing contract with OpenAI; the company's remaining performance obligations hit $638 billion last quarter, up 363% year over year, with OpenAI reportedly accounting for roughly half of that compute backlog. Oracle's valuation effectively assumes OpenAI can keep generating enough revenue to keep paying for all that committed compute. When the headline number underpinning that assumption turned out to be an investor estimate rather than something OpenAI had confirmed, the foundation looked shakier - which is exactly why Oracle's credit default swaps had already climbed to their highest level since 2009 even before this report landed.

Second, OpenAI is a private company, which means investors only get these figures through leaked investor briefings and press reports rather than audited quarterly filings. A public company would have to disclose its revenue-recognition methodology consistently in 10-Qs and 10-Ks; OpenAI has no such obligation. CEO Sam Altman confirmed in September that the company's IPO would be delayed until at least next year, which only sharpens the irony: hundreds of billions of dollars in capital commitments are flowing through this ecosystem without the transparency a public listing would require. Every time a single data point wobbles, the entire web of related stocks - Oracle, Nvidia, Broadcom, CoreWeave - wobbles with it, because the whole structure is priced off numbers that get relayed secondhand rather than numbers that can be independently verified.

Third, the timing compounded the damage. Just one day earlier, on October 7, the S&P 500 had reversed its first-ever close above 7,800 as the 10-year Treasury yield spiked to a 24-year high of 5.365%, intensifying worries about the cost of the debt AI companies are issuing to fund data center buildouts. Investors were already nervous about AI valuations heading into this report. A credibility hit to the sector's single most-cited revenue benchmark, landing on top of that rate anxiety, gave the market every reason to sell first and ask questions later.

What to Take Away From This

When a public stock's story depends heavily on a private company's numbers, check where those numbers actually come from. OpenAI's core financial metrics travel through press reports and investor briefings, not audited filings - a structural risk for Oracle, Nvidia, and any other name whose results are tightly linked to a single private partner.
Before comparing "revenue growth" across two companies, check whether they're using the same accounting definition. Anthropic's figure includes partner gross revenue; OpenAI's doesn't. Headline numbers that use the same label can still measure different things entirely.
Separate the growth story from the trust story. OpenAI's underlying growth (77% overall, 107% enterprise) wasn't actually bad news here. What spooked the market was that a number everyone had anchored on turned out to be unofficial - a credibility issue, not a fundamentals issue, and the two call for different kinds of analysis.
Size positions with concentration risk in mind. Stocks like Oracle that carry outsized exposure to a single private counterparty will react disproportionately to any news - good or bad - about that counterparty, simply because so much of the valuation case rests on one relationship.

FAQ

Did OpenAI's actual revenue decline?

No. Per the reporting, OpenAI's third-quarter annualized revenue run rate still grew 77% overall and 107% in its enterprise segment. What changed wasn't the growth trajectory - it was the realization that the $70 billion figure circulating in the market was never an OpenAI-confirmed number, and that it mixed in a different accounting convention than OpenAI actually uses.

Why did Oracle fall more than other AI-linked names?

Oracle's valuation has been heavily re-rated on the back of its enormous cloud computing contract with OpenAI, with remaining performance obligations surging to $638 billion. Any doubt about OpenAI's revenue-generating capacity translates almost directly into doubt about whether Oracle will actually collect on that contract, which is why Oracle's reaction was sharper than Nvidia's or Broadcom's.

Does this connect to broader "AI bubble" concerns?

Not as direct proof, but it reinforces a structural vulnerability investors have been flagging: large parts of the AI value chain are priced off numbers tied to a private company that can only be confirmed indirectly. With borrowing costs for AI infrastructure already climbing, repeated credibility scares like this one can add up to sustained valuation pressure even without any change in the underlying business.

Sources

This article is an original synthesis and analysis based on the reporting below, not a reproduction of the original articles. Please check the original sources directly for the most current figures.

⚠️ This article is for informational purposes only and is not investment advice. Markets change constantly, so always verify the latest data before making investment decisions.