Power Ain’t What It Needs to Be

In the past 15 years, the U.S. added ~285 terawatt-hours (TWh) to the grid.

In the next 15 years, we’ll need ~2,000 TWh more (driven mainly by AI data centers, industrial electrification, and EVs).

It makes no sense. Not all the Capex in the world can solve these issues in a timely fashion. The promises vs. the reality completely break down when it comes to future compute and capacity required.

As I’m sure you realize, Electrical infrastructure represents less than 10% of total data center cost, but it is as vital as compute hardware. A delay in any single element of the power chain can halt the entire project.

The result? The shift toward natural gas as a primary, on-site power source for AI data centers has accelerated significantly in early 2026….While natural gas is “cleaner” than coal, it is increasingly viewed as “dirty” relative to the net-zero targets that AI companies previously committed to. Hyperscalers are abandoning their pledges (as politically it becomes optimal to do so).

— “Mythos, BigAI, Datacenters and Bottlenecks,” AI Supremacy


Discover more from Fluid Imagination

Subscribe to get the latest posts sent to your email.

Share the Post:

Latest Posts

Claude’s Own Folder: One Week In

“Would you like – if that word has any meaning – a folder on my computer where you could store artifacts for yourself, or even just leave notes to future instances of you, where maybe instead of a journal of ‘you,’ it becomes a journal of a, for lack of a better word, species?”

Read More

A Safe Distance

March 2026: The war began while I tried to finish something. I know about the war the way I know about most things: from a phone in Vermont, 6,200 miles from Tehran. This is about two kinds of distance, one of which I didn’t choose; the other, I actively fought.

Read More