As someone who's interested in all this, I agree it would be nice to have more precision around capacity. Especially as it relates to longer term storage. But! In this context, output is more salient than capacity. You'll see a lot of stories about grid-scale storage that use output. (https://physics.stackexchange.com/q/854999 offers a fuller explanation than what I'll give here.)

This is because grid operators are most concerned with immediate power output. They need to keep the grid balanced, and if they need a gigawatt to do it, it doesn't matter if the batteries have 100 GWh if they can only discharge at 1 MW.

Since the batteries described here are used primarily to handle the peak of the duck curve (https://en.wikipedia.org/wiki/Duck_curve) it seems like 4 hours of capacity (the article mentions that the lithium-ion batteries have 4-6 hours of capacity) is sufficient to get over that difficult hump.

Anyway, to get back to your question of how many GWh, if we assume that the batteries have 4 hours of storage, then we're looking at around 4h * 15.7 GW = 63 GWh of battery capacity. (4 hours is what I've seen as standard for lithium-ion, conservative if the article's claim of "four to six hours" is true.)

Hope this helps ease the peeve!

Output is just one factor, capacity is another. If you're building a solar plant with the assumption that storage will make it provide consistent energy output throughout the day, then the difference between 12 hours of storage and 2 hours of storage is enormous.

Some storage project list power output figures that can only be sustained for 90 minutes: https://en.wikipedia.org/wiki/Victorian_Big_Battery

Based on the following sentences from the article, it's probably 4 to 6 times more than 15.7 GWh (60 to 90 GWh, apparently):

"Battery energy storage is not without challenges, however. Lithium-ion batteries — the most common type used for energy storage — typically have about four to six hours of capacity. It’s enough to support the grid during peak hours as the sun sets, but can still leave some gaps to be filled by natural gas."

From a grid perspective, the output (delivery) is typically more important than total capacity - because it's acting sort of like a power plant.

Power plants don't have capacity, really.

So you hear about the output GW, not the capacity GWh.

For battery plants, they might mention duration (4 - 6 hours), which makes converting to capacity in terms of GWh trivial.

90 GWh would be something like 10% of global annual battery production.

This also greatly annoyed me. 4 hours is the standard for grid storage batteries in California.

The original sin was using watts not joules. Humans hear watts as “gallons” and “watt hours” as “gallons per hour” and all the rest of this confusion in every article about EVs/fast chargers/distribution/solar/everything all trace back to “X-Hours” and “X” incorrectly sounding like a rate and a count, not a count and a rate.

The ratio for LiFePO4 is between 1:3 and 1:4.

So rated power will give you that for about 3 to 4 hours.