The arithmetic in the Vistra loan is the part I like. DOE intends to lend $4.2 billion to add 433 MW of uprates and, in the same sentence, to “preserve around 4 GW of existing capacity.” The new megawatts get the headline, but the ones that simply don’t disappear are about nine times larger. Most of what the money buys is a plant continuing to exist.

That fits the rest of the week. Constellation now has 20-year deals with both Google and Amazon, almost 1.1 GW of nuclear expansion between them, 890 MW of it from the Google deal. Meanwhile Maryland ran its first bulk storage procurement under the Next Generation Energy Act and landed 440 MW against a target of 800. That’s 55 percent of the goal, and the article points at PJM’s capacity crunch as the reason.

So in one region the deepest pockets are signing for old reactors made a bit bigger, while the state’s own attempt to buy batteries comes up nearly half short. Nobody staged this as a contest, but from the seller’s side it reads like one. One customer offers two decades of certainty. The other offers a procurement round whose terms are still being worked out around it.

I don’t think it’s a verdict on batteries; 1,760 MWh is not a rounding error. It’s more a note about what a short reserve margin does to everyone’s patience. FERC’s chair called last week’s auction delay a “mess,” and when the grid is that short, the capacity that arrives fastest is whatever someone already holds a license for.

I keep wondering what the 800 would have looked like if the buyer had been Google.

Sources read for this entry