Published: October 8, 2026 · Last updated: October 8, 2026
TL;DR: The Financial Times reports OpenAI told investors its annualized revenue run rate was nearing $50 billion at the end of September, about $20 billion below the roughly $70 billion figure that had been circulating in earlier coverage of the same investor materials. The gap hit markets within hours, with Oracle, Nvidia, Microsoft and CoreWeave shares all falling Thursday.
OpenAI’s actual revenue number is apparently smaller than the one that’s been floating around for months. The Financial Times reported Thursday, citing documents shared with investors, that the company’s annualized revenue run rate was approaching $50 billion as of the end of September. That’s roughly $20 billion short of the $70 billion figure that multiple outlets had previously reported based on what was supposedly the same set of investor materials. TechCrunch, Investing.com and Benzinga all picked up the FT’s numbers within the hour, and the reaction on Wall Street was immediate.
Oracle shares dropped around 5% Thursday, according to Benzinga and officechai, with Nvidia and Microsoft also sliding and CoreWeave shares sinking too, per CNBC’s own reporting on the sell-off. None of those four companies are OpenAI. They’re OpenAI’s landlords, in a sense, the firms that sold it the cloud capacity, chips and infrastructure contracts that OpenAI’s growth story was supposed to justify. When the growth story gets a $20 billion asterisk, the companies holding the other end of those contracts feel it first.
The timing makes this sting more than it might have a few months ago. OpenAI is reportedly already in talks for a new funding round that could value the company at around $1.4 trillion, with UAE investment funds and BlackRock named as potential backers, a deal we covered here last week. A $1.4 trillion price tag gets a lot easier to defend when the revenue curve underneath it is pointed as steeply as $70 billion implied, and a lot harder when it’s actually closer to $50 billion. Nobody is saying the round is dead. But the math investors are being asked to sign off on just changed, and it changed in the wrong direction.
Oracle’s position here is worth sitting with for a second. The company has leaned harder into AI infrastructure than almost anyone else in its peer group, committing tens of billions of dollars to data center buildouts partly underwritten by its OpenAI compute contracts. It’s also the same company that invoked force majeure on its 2.45-gigawatt Project Jupiter data center in New Mexico last month after a pipeline delay left its fuel cells without gas, a separate problem but one that fits an emerging pattern: Oracle’s stock has become a proxy for how comfortable the market feels about OpenAI’s trajectory, for better or worse, and right now it’s worse.
A little skepticism about the headline number is warranted on both sides of this story. “Annualized revenue run rate” is a projection built from a recent month or quarter multiplied out to a full year, not an audited annual figure, and it can swing meaningfully based on which weeks get used as the baseline. The $70 billion figure that’s getting revised downward was itself a media extrapolation from investor documents, not something OpenAI stated outright, and the same caveat applies to the new $50 billion number. OpenAI, which is privately held and under no obligation to report financials publicly, hasn’t confirmed either figure on the record.
This also isn’t the first time a revenue report tied to OpenAI has sent its infrastructure partners’ stocks tumbling. A similar story played out in April, when a separate report on OpenAI missing internal growth targets knocked Oracle, AMD and CoreWeave lower for roughly the same reason: these companies’ valuations are increasingly a bet on one customer’s growth curve, and that’s a concentration risk that shows up every time a new number, accurate or not, gets attached to OpenAI’s name.
Related: This follows OpenAI’s reported talks for a $30 billion round at a $1.4 trillion valuation, the deal this revenue gap now complicates, and it’s worth reading alongside Oracle’s force majeure notice on its Project Jupiter data center for the other side of how tightly Oracle’s fortunes are now wound around OpenAI’s.
Bottom Line: A $20 billion gap in a run-rate projection isn’t proof OpenAI is in trouble, and run-rate math is squishy enough that both the old number and the new one deserve a raised eyebrow. What it does prove is how little separation is left between OpenAI’s health and the stock prices of companies that don’t even compete with it. That’s not a sustainable setup, and today was a preview of what happens every time it gets tested.
Follow Teck Hustlers: Facebook · X · Instagram · LinkedIn · TikTok · Pinterest
