AI Unplugged: When the Bill Comes Due

I'm starting this series of articles with what I think is a pivotal turning point in the rapid development of AI. And it starts with Sam Altman of OpenAI asking for $7 trillion .

Seven. Trillion. Dollars.

That's a lot of money. Like, mind-boggling amounts. But it's because the AI bill is coming due. As Altman himself said , “the world needs more ai infrastructure--fab capacity, energy, datacenters, etc--than people are currently planning to build ... building massive-scale AI infrastructure, and a resilient supply chain, is crucial to economic competitiveness."

If you unpack that statement, Altman's not just talking about the world needing this investment, but that governments need to pay attention too (those two words, "economic" and "competitiveness" are aimed at them).

But Why?

Why all this money?

Scott Alexander at Astral Codex Ten has one of the best answers I've read , and it goes like this:

Powering AI is challenging enough once it's up and running, but training it requires an enormous surge of capacity, as the AI isn't making any return of investment during that time. As Alexander points out, we are rapidly approaching the upper limits of computing and plain old regular power to keep all this going. The equivalent of ChatGPT-5 is an enormous amount of power (1% of the entire world), and it gets worse from there for GPT-6 and 7. We're going to need chips, we're going to need power plants, we're going to need infrastructure -- everything Altman himself said in the above tweet. And it's going to cost trillions of dollars.

He's probably not going to get it. But in a lot of ways, governments are increasingly seeing AI as a form of soft power, a weapon that nations can wield for their own interest and against competitors. If you look at AI investment like nuclear investment, $756 billion (the amount the U.S. currently invests in nuclear forces ) seems about right. Let's be optimistic and round up to $1 trillion between 2023 to 2032. Still not enough. But don't worry, Altman's talking to the United Arab Emirates, who might well have that kind of ready cash.

Who Pays?

And what if they don't? The real challenge is this: if we don't have the resources we need to fund AI development, but companies and governments plan to make it a priority (because the opportunity for profit is unbelievably immense), who pays that $6 trillion dollar shortfall in resources, energy, and infrastructure?

The answer is going to define our future. Buckle up.

Please Note: The views and opinions expressed here are solely my own and do not necessarily represent those of my employer or any other organization.