Trump's energy policy is incoherent and self-defeating

You may have heard that the cost of electricity is increasing in America. It’s true! Even as inflation has moderated in general, electricity prices have surged:

It’s important to put this increase in context. When we measure electricity prices relative to how much people earn, we find that the recent rise in prices is pretty modest. In fact, so far it hasn’t even canceled out the big drop in electricity prices in the late 2010s:

But rising costs are rising costs, and it would be a bad idea to wait until things get really bad before we address the problem.

What’s more, this is the exact opposite of the direction that electricity costs ought to be going in. We’re in the middle of a miraculous revolution in energy technology. The cost of solar power and battery storage has absolutely plunged in recent years, to the point where developing countries like India are choosing renewables over coal simply because it’s cheaper to do so. Energy prices should be going down, not up. And in China prices are going down, in fact, despite an explosion in demand. Why is America different?

Everyone agrees that there are a bunch of factors at work. For example, here’s Wired:

There are several dynamics driving the current power price spike. Rising electricity demand, volatile fuel prices, inflation, tariffs, a slowdown in transmission line construction, and long delays in adding new generators to the power grid are all conspiring to create more expensive utility bills.

The Wall Street Journal also cites the increasing cost of rebuilding power infrastructure after natural disasters.

But one factor that lots of people cite is the AI boom. AI runs on computing power (or “compute”, as they say). You need compute to train models, and you also need compute to make AI “think” about the answers it’s giving. Scaling up this latter type — which is called “inference compute” — is now the most important way that AI companies are improving their models over time. And compute requires energy, because you’re running electricity through a bunch of chips in some data center somewhere. Every time you ask AI a question, it’s using some noticeable amount of electricity.

People go back and forth about how much the increased electricity demand from AI is already affecting electricity prices. A recent report from Lawrence Berkeley National Laboratory argues that AI hasn’t affected electricity prices yet, pointing out that states that saw greater electricity demand growth actually had falling prices over the past few years:

But Bloomberg Technology disagrees. Their analysis looks at smaller geographic areas, and considers wholesale rather than retail prices, and argues that data centers are already having big local effects:

A Bloomberg News analysis of wholesale electricity prices for tens of thousands of locations across the country reveals the effects of the AI boom on the power market with unprecedented granularity. The locations and prices were tracked and aggregated monthly by Grid Status, an energy data analytics platform. Bloomberg analyzed this data in relation to data center locations, from DC Byte, and found that electricity now costs as much as 267% more for a single month than it did five years ago in areas located near significant data center activity.

As the country builds a lot more data centers, we can expect this effect to filter through to