#OpenAIAnnualRecurringRevenueNears$70B

OpenAI approaching a $70 billion annualized revenue run rate is the kind of number that makes the AI valuation debate much more interesting.

According to Axios, OpenAI's annualized revenue run rate is now nearing $70 billion, up more than 70% since the beginning of Q3, while enterprise sales have more than doubled since July. On the consumer side, the company reportedly added more revenue during Q3 than it added throughout all of 2025.

That is not a small change.

What stands out to me is not just the $70 billion headline. It's the speed of the growth and, especially, the acceleration in enterprise demand.

Consumer AI created the initial explosion, but enterprise adoption is where the business model could become much more interesting. Companies are increasingly willing to pay for AI because they are trying to integrate it into coding, research, customer service, productivity and internal workflows. If enterprise revenue continues scaling at this pace, OpenAI starts looking less like a consumer software story and more like a large-scale technology infrastructure and enterprise platform.

But this is where I would slow down before jumping straight from $70B revenue run rate to “therefore the valuation is justified.”

ARR is a run-rate metric. It annualizes current revenue performance; it is not the same thing as OpenAI actually booking $70 billion of revenue in a completed year. Reuters also notes that annualized revenue run rates can be misleading because they extrapolate a shorter period of sales across a full year.

So the next question isn't simply:

“How fast is OpenAI growing?”

It's:

“How much of that growth can eventually become durable, profitable cash flow?”

That's a much harder question.

AI companies have an unusual cost structure. Every additional user and every additional enterprise workload can generate revenue, but it also creates enormous demand for computing power, data centers, chips, networking and electricity.

And OpenAI is operating in one of the most capital-intensive technology races we have seen.

That's why I think the revenue side and the spending side have to be analyzed together.

The latest reporting around OpenAI's potential financing discussions points to a valuation of roughly $1.4 trillion, although those discussions are preliminary and the terms can change. Reuters reported that the company is considering raising at least $30 billion, while its previous March funding round valued it at about $852 billion. OpenAI's CEO has also said the company would not go public in 2026.

That gives us an interesting situation.

The market is putting an enormous valuation on a company whose revenue is growing extremely quickly — but investors also need to believe that the company can scale its economics alongside that revenue.

For me, that's the real AI valuation debate.

If OpenAI can continue increasing enterprise revenue, maintain strong demand for its models and eventually improve the economics of serving those workloads, then today's enormous revenue growth could become the foundation for an even larger business.

But if revenue grows rapidly while compute and infrastructure costs grow just as aggressively, then the headline revenue number alone tells us much less about the eventual economics.

This is also why the enterprise number catches my attention.

Enterprise sales more than doubling since July suggests that AI demand isn't only coming from individual consumers experimenting with ChatGPT. Businesses are increasingly becoming paying customers, and that could make revenue more recurring and embedded in corporate workflows.

At the same time, competition is moving extremely quickly.

The AI market is no longer a one-company story. Other frontier-model companies are also seeing enormous revenue growth, while investors are pouring capital into the infrastructure required to support these models.

That means the valuation race isn't just about who has the biggest model.

It's about who can turn model capability into recurring revenue, recurring revenue into margins, and margins into sustainable free cash flow.

That's the part I want to see over the next few years.

So when I look at the nearly $70B annualized revenue figure, I don't see it as proof that any particular valuation is automatically cheap or expensive.

I see it as evidence that the commercial side of AI is becoming much larger, much faster than many people expected.

The next phase is going to be more difficult.

Can that growth continue?

Can enterprise adoption keep accelerating?

Can AI inference costs fall fast enough?

Can infrastructure spending remain economically sustainable?

And ultimately, can OpenAI turn this extraordinary revenue growth into durable profits?

That's where the real valuation test begins.

The $70B number is impressive.

The 70%+ growth is even more interesting.

And the more-than-doubling of enterprise sales since July may be the most important part of the story.

But revenue growth is only one side of the equation.

AI has already proved that people will pay for intelligence.

Now the market needs to find out how much sustainable profit can be built behind that demand.