#OpenAIAnnualRecurringRevenueNears$70B

OpenAI getting close to a $70 billion annualized revenue run rate is a huge number. But honestly, the revenue figure itself is not what interests me most.

What matters is what happens underneath it.

OpenAI's reported revenue pace has accelerated sharply, with enterprise demand becoming a much bigger part of the business. That tells me the AI market is moving into a different phase. A few years ago, the biggest question was whether people would actually use generative AI. Then it became whether businesses would pay for it. Now the question is whether those payments can turn into a durable, profitable business at massive scale.

And I think that distinction is extremely important.

A $70 billion annualized run rate does not mean OpenAI has already generated $70 billion in revenue. It is simply the current revenue pace extrapolated over a year. If growth slows, the annualized number changes. So I wouldn't look at $70B and immediately say the company is worth a certain amount.

I'd look at what kind of revenue is producing that number.

Enterprise adoption is particularly interesting to me because business customers can become deeply embedded in AI workflows. Coding, research, customer support, internal knowledge, automation and productivity are very different from someone occasionally opening an AI chatbot. Once a company builds AI into daily operations, switching becomes harder and the potential value of the service becomes much larger.

But there is a problem hiding underneath all this growth: compute.

AI is not ordinary software. More customers mean more inference. More inference means more GPUs, memory, networking, data centers and electricity. If revenue is growing 70%, but the infrastructure required to generate that revenue is also exploding, then headline growth alone doesn't tell us enough.

That's why my focus is shifting from “How fast is OpenAI growing?” to “How efficiently can OpenAI scale?”

If model efficiency improves, hardware utilization gets better and inference costs fall while revenue continues climbing, the economics could change dramatically. Revenue growth would start translating into much stronger margins.

But if every major jump in revenue requires another enormous wave of infrastructure spending, then investors will eventually care much more about capital intensity and free cash flow.

This is also why the valuation conversation is getting harder. A business approaching this revenue scale can justify enormous expectations, but the higher the valuation becomes, the more execution the market is demanding in return.

My take is that AI has already passed the first test: people are willing to pay for intelligence.

Enterprise adoption is helping prove the second test: businesses are willing to make AI part of their operations.

Now comes the hardest test:

Can AI companies turn that enormous demand into sustainable profits without infrastructure costs growing at the same speed?

That is the number I would watch behind the $70B headline.

Revenue tells us how big the AI market is becoming.

Margins and free cash flow will tell us how valuable that growth really is.

@Gate_Square @GateSquare