OpenAI Is at $50 Billion a Year and Wants $30 Billion More
OpenAI just gave investors its latest numbers, and they say two things at once. The company is growing very fast. It is also smaller than a lot of people thought.
According to the Financial Times, OpenAI's annualized revenue rate was close to $50 billion at the end of September. At the same time, it is in talks to raise at least $30 billion in new money. Here's what's going on.
The Headline Numbers
Start with the growth, because it is real. OpenAI's total revenue in the third quarter was up 77%, and the company says enterprise customers drove most of that.
Then there's the raise. OpenAI is reportedly targeting a pre-money valuation of $1.4 trillion. That would put it among the most valuable companies on earth, public or private. It would also make it one of the largest private funding rounds ever attempted.
So far, so normal for AI in 2026: big growth, bigger valuation.
The $20 Billion "Gap"
The surprise was the size. Outside estimates had OpenAI's run rate closer to $70 billion. The $50 billion figure is roughly $20 billion short.
But the shortfall is mostly about how revenue gets counted, not missing customers. When Anthropic sells its models through partners like AWS and Google Cloud, it books those sales as its own revenue. OpenAI leaves equivalent partner sales out of its own figure.
Both methods follow US accounting rules (GAAP). They just answer different questions. One says "how much money flowed through our models." The other says "how much money landed with us." If you compare the two companies line by line without adjusting, OpenAI looks smaller than it would under Anthropic's method.
Why the Market Flinched Anyway
Accounting footnotes or not, investors reacted. After the FT report, the Nasdaq 100 closed down 1.4%. Nvidia fell 2.9%. Oracle, which has tied a lot of its future to building data centers for OpenAI, dropped 5.5%.
That reaction tells you something. A huge share of the AI buildout (chips, data centers, power contracts) is priced on the assumption that frontier labs keep growing fast enough to pay for it. When the biggest lab's top line comes in below what people pencilled in, everybody downstream gets nervous, even if the reason is a definition.
The Real Story: Enterprise
The most useful detail in the report is the one that got the least attention. Enterprise drove the quarter.
Consumer ChatGPT subscriptions made OpenAI famous. But the money is increasingly coming from companies paying to put AI to work: agents handling support tickets, coding assistants, internal tools wired into business systems. That is the same shift happening across the industry. The value is moving from "chat with an AI" toward "an AI that does tasks for you, every day, without being asked twice."
That also explains why the valuation can stay so high while the revenue number disappoints. Investors aren't paying for today's chatbot. They are betting that AI becomes ongoing labor that businesses pay for every month.
What to Watch Next
A few things will tell you whether this was a blip or a turning point:
- Whether the $30 billion round closes at $1.4 trillion. A lower valuation would be a much louder signal than one FT story.
- Whether labs start reporting revenue the same way. Until they do, comparisons between OpenAI and Anthropic will keep producing confusing headlines.
- Enterprise growth rates. If business customers keep growing at the pace that drove a 77% quarter, the accounting noise won't matter much.
The Bigger Picture
Strip away the valuation drama and the lesson is simple. The AI that earns money is the AI that gets work done. Businesses are paying for agents that remember context, use real tools, and keep going without someone typing every step.
That's the idea behind OpenClaw. Your agent runs persistently, keeps its memory between sessions, and takes care of tasks on its own schedule. You don't need a trillion-dollar budget to get that kind of help, just an agent that's always on and actually knows what you're working on.