There’s a lot of focus right now on model performance who’s ahead, who’s scaling faster, what breakthroughs are coming next.

But when you zoom out and look at the infrastructure layer, a different constraint starts to show up: energy.

Recent projections around AI infrastructure don’t just point to steady growth in electricity demand they suggest something much sharper. Demand isn’t increasing evenly; it’s concentrating in specific regions where data centers are being rapidly deployed.

Those regions tend to share a few characteristics:Those regions tend to share a few characteristics:

√ relatively cheap and stable power

√ available grid capacity (or the ability to expand it quickly)

√ regulatory environments that allow fast buildouts

Which makes sense operationally. But structurally, it creates imbalance.

Because if access to large scale computation depends on access to large scale energy, then the map of AI capability starts to overlap heavily with the map of energy infrastructure.

And that raises a bigger question:

👉 Does energy become the limiting factor for who can participate meaningfully in the AI economy?

Not just in terms of innovation, but in terms of ownership and control.

If that’s the case, then this isn’t just a technical constraint it’s a geopolitical one.

And it might end up mattering more than model architecture itself.