A single figure was enough on Thursday, October 8, 2026, to shake the companies most tied to the artificial intelligence boom. The Financial Times reported that OpenAI had told its investors that its annualized revenue was approaching $50 billion at the end of September. The figure contrasted with the roughly $70 billion that had circulated in the media weeks earlier, and the market reaction was immediate.
CNBC confirmed the figure of about $50 billion. According to a person familiar with the matter cited by the network, the $68 billion figure that circulated at the end of September included gross revenue from OpenAI’s partners, a method meant to make comparisons with its main rival, Anthropic, easier. The new figure, by contrast, is closer to the company’s own net revenue. OpenAI declined to comment to the Financial Times.
The source of the confusion, according to the British paper, lies in adjustments made by some investors. Anthropic counts the full value of sales made through cloud providers, while OpenAI did not count them the same way. To compare the two companies, some raised OpenAI’s figure, which led to reports that it had reached $40 billion annualized in August. When OpenAI later said its revenue had grown by more than 70%, that inflated starting point produced the figure of about $70 billion.
It is worth clarifying what the figure means. Annualized revenue projects the sales pace of a shorter period, usually one month, over a full year. It is a metric criticized as imprecise, but investors still use it as the clearest reading of artificial intelligence demand at OpenAI and Anthropic. The figure being $20 billion lower does not mean the company lost that money, but that the earlier comparison started from different accounting bases.
Even so, the market reacted as if it were bad news. According to SiliconANGLE, Oracle, which has multibillion-dollar contracts with OpenAI, fell 5.5%; Intel dropped 5.3% and Nvidia 2.9%. CoreWeave, a specialist in cloud infrastructure built on graphics processors, lost 8%, while AMD and Broadcom fell about 4% and server maker Super Micro Computer almost 5%. The Nasdaq 100 lost about 1.4%, according to Bloomberg.
The site Riskstock calculated that the Nasdaq closed down 1.25% at 27,193 points, its worst session since mid-August, while the S&P 500 slipped 0.47%. Several analysts noted that rising oil prices and U.S. Treasury yields also weighed on the session, so not all of the decline can be attributed to OpenAI.
The reason companies like Oracle and CoreWeave suffer more is their direct exposure. Both have signed very large contracts to supply computing capacity to OpenAI and have borrowed to build data centers for it. If their customer’s revenue is lower than thought, investors ask more insistently whether those contracts can be paid in full on the expected schedule.
OpenAI’s growth, in any case, is still very fast. According to the person cited by CNBC, the investor presentation highlighted 77% growth in the total revenue run rate during the third quarter and 107% in its enterprise business. To put the figures in perspective, OpenAI recorded about $13 billion in revenue for all of 2025. Bloomberg also reported that the company expects to reach or exceed $70 billion annualized by the end of this year, driven mainly by corporate customers.
The episode comes at a delicate moment. According to the Financial Times, OpenAI is negotiating a private round that could value it at around $1.4 trillion, and Bloomberg reported talks to raise $30 billion or more at that valuation. The company confidentially filed to go public in June but, according to Bloomberg, has pushed the listing back to at least next year. Anthropic, which told investors its revenue run rate reached $65 billion at the end of July, could debut on the stock market in November.
Some observers called the sell-off an overreaction, since it involved a comparison between figures that were not equivalent. But the episode leaves several lessons. The first is the fragility of the artificial intelligence growth narrative: an accounting correction at a private company was enough to trigger sharp stock market declines at some of the largest technology companies within hours. The second is the lack of transparency at companies that, despite their enormous influence on markets, still do not publish audited accounts.
The third lesson concerns the comparison between OpenAI and Anthropic, which has become a barometer for the sector. As long as the two companies do not apply consistent criteria to count their revenue, any direct comparison calls for caution. Their future stock market listings will force both to publish financial statements under common rules, which should clarify their real size.
For infrastructure investors, the underlying question is whether demand for artificial intelligence will grow fast enough to justify the hundreds of billions being invested in data centers, chips and energy. OpenAI’s figures still show very strong growth, but the market has made clear that growing is not enough: companies also have to meet the expectations that have built up.
In short, the session of October 8 was a reminder that enthusiasm for artificial intelligence rests largely on partial figures and third-party estimates. OpenAI remains one of the fastest-growing businesses in history, but the value of many listed companies depends on that trajectory continuing. Until the sector’s major companies publish complete and comparable accounts, episodes like this one may happen again.
