Who Pays for Electricity?

March 9, 2026

green-dino_thumbAnother dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.

The cost of electricity is an interesting topic. Figuring out power demand and the cost of the infrastructure requires research, math, experience, and knowledge of technical options. I worked years ago with a utility rate expert. He was a specialist, and he was, no pun intended, in demand.

I thought about the easy assumption that power is just available. Buy a gizmo, plug it in, and it turns on. The electric bill arrives, and it is relatively predictable. Rates typically don’t gyrate chaotically. When my family lived in Brazil, the power would go off but the “cost” of the electricity was reasonably predictable. What happens when one assumes that a data center can be built and plugged in? The electricity is there and the bill arrives. No big deal.

image

When the batteries go dead, recharging may be a challenge. Thanks, Venice.ai. Good enough.

Nope. Big deal. Utility demand forecasting is tricky when nothing wobbles too far from the forecasters’ and analysts’ projections. Plugging in numerous data centers means that the complex machinery for power generation has to be refigured. Behind the scenes of what is to most people a power generation company are many moving parts. Out of sight and out of mind does not mean easy to operate and trivial to scale.

I thought about this assumption that electricity is just “there” when I read “Tech Giants Sign Energy Pledge at White House Ahead of Midterms.” The operative word is “pledge.” Promising to pay for electricity required to run the data centers some companies are building, trying to build, planning to build, and hoping to operate is just that a promise. A promise is not cash. Furthermore, I am not sure the “tech giants” know the cost of the electricity their AI factories will require.

The BBC article “Tech Firms Pledge to Pay for AI data Center Power Costs. But Will They?” cuts to the core of the electricity problem. The write up states:

The tech pledge may be difficult to enforce, said John Quigley, a senior fellow at the Kleinman Center for Energy Policy at the University of Pennsylvania. He cited the multiple layers of government, grid managers and electricity regulators involved in power projects.

Saying is one thing. Doing is another.

When it comes to power generation, fast talk may provide superficial reassurance to some people. The shift from a pledge to forking over money may not be the same as putting one’s luxury car in gear and allowing the smart self driving vehicle to whisk the passengers to an off site face to face meeting. How is that working out by the way?

Several observations:

  1. Rolling out cost effective power generation takes time, specialty manufacturing, and planning. One does not ring up a company and have a modular nuclear generating machine dropped off at the loading dock.
  2. Creating electricity and getting it where it needs to be are shotgun marriages of old-fashioned technology, advanced materials, and infrastructure both physical and digital. The costs to ensure a happy, stable union are usually steep and difficult to project with accuracy.
  3. People living near a new or expanded power generation facility, infrastructure, or a big data center might throw a wrench in the works. What few people realize is that power generation is one of the more vulnerable “it’s just there” services. Angry consumers can create some tricky challenges easily and quickly.

Net net: Pledges are not power. Pledges are words. Power generation makes modern life possible, yet few whiz kids know anything about the business. When the power goes away as it did where my family lived in Brazil, one needs a Plan B. A pledge is not a Plan B. A pledge is public relations.

Stephen E Arnold, March 9, 2026

Comments

Got something to say?





  • Archives

  • Recent Posts

  • Meta