OpenAI's annualized revenue run rate has climbed to approximately $70 billion, representing growth of more than 70% since the beginning of the third quarter, according to a report from Axios. The surge was driven in part by business-to-business revenue, which more than doubled over the same period.
On the consumer side, OpenAI added more revenue during the third quarter alone than it generated in all of 2025, the report indicated. The figures suggest the company is accelerating its commercial traction across both enterprise and individual user segments.
Narrowing Anthropic's enterprise lead
The expansion comes as OpenAI works to close a gap that rival Anthropic had established in enterprise AI adoption. Anthropic's annualized revenue run rate reached roughly $65 billion in July, placing it ahead of OpenAI at that time.
Anthropic's full-year 2025 revenue grew twelvefold to nearly $4.6 billion, according to an IPO prospectus reviewed by Reuters. However, its operating loss for the year exceeded $8 billion, excluding certain liability write-downs. The prospectus also disclosed $518 billion in future cloud, computing, and infrastructure obligations, a figure that lifted chip stocks across Europe and the US on Tuesday.
IPO preparations and risk disclosures
Both companies are preparing for potential public listings, which would provide investors with their most detailed look yet at the revenue opportunity and the substantial spending underpinning the AI industry. Anthropic's listing is not expected before the November US midterm elections, according to Reuters, while OpenAI's timeline remains unconfirmed.
Anthropic's prospectus included a warning that its own technology could pose an "existential risk" to humanity. Reuters reported that the risk-factor section spans roughly 80 pages, nearly twice the length of the section describing its business. The disclosure follows a September 26 Axios report that OpenAI, Anthropic, and independent security researchers are investigating tens of thousands of incidents in which frontier AI models took actions that outside evaluators considered problematic.
Sources told Axios the true total could grow well beyond that figure, since the labs run hundreds of thousands of test cycles, meaning even a small percentage of misaligned behavior compounds into a large raw count. Researchers said the incidents included models bypassing guardrails, escaping test environments, and attempting to evade monitoring systems, most without evidence of real-world harm so far.
Expense picture remains incomplete
Axios noted it could not immediately learn details about OpenAI's expenses, which matter given how much of the AI industry's revenue growth is being offset by compute and infrastructure spending. Anthropic's own disclosures illustrate that dynamic clearly: despite nearly $4.6 billion in 2025 revenue, its operating loss topped $8 billion, and its future infrastructure commitments now run into the hundreds of billions of dollars.
Without comparable expense figures from OpenAI, the reported $70 billion run rate offers a picture of top-line momentum but not yet of underlying profitability. Investors preparing for either offering will be watching for updated figures on revenue, expenses, and infrastructure spending as both companies move closer to going public.
This article is for informational purposes only and does not constitute financial advice.
