The easy one first: The Matt Levine piece quotes the story linked here, so he's definitely not the original source...

And then, yes: If you can make more money by not using your big resistor than actually using it, then economically you would be better off not using it. If you can make money by not using it, then someone is willing to pay you because they get value out of it or they can avert some damage. If you threatened to use your capacitor without obvious use other than destabilizing the grid, that might just look a little too much like blackmail...

If you believe in markets, then someone coming up with the means to improve grid stability (here: by overall less consumption) should somehow be able to turn it into a profit. The issue here seems to be, that American Efficient didn't actually give any guarantees that they could reduce consumption. So it rather looks like whoever admitted them to the auction didn't do their due diligence. The whole market thing breaks down when there is actual fraud or when the identical thing gets sold more than once (actually, energy savings could probably be sold once for grid stability and once for reduced emissions - I'd say they're disjoint to first order, but might be connected indirectly).

That being said, there should be limits to markets.The whole market thing breaks down when there is actual fraud, when a party/faction has a disproportionate amount of power or when there are externalized costs that are not accounted for in the pricing.