You are both very much missing the point. Almost all power usage should be directly balanced by generation. When demand goes very high, prices go very high, all available generation runs at whatever level it can and demand reduces, through reducing most commercial use, until things match. When generation goes high, prices reduce, all available demand runs and most generation that can be forced off the grid is (typically with the exception of nuclear and some home solar, though that is voltage limited).
What the storage does is takes the lowest possible point of price, uses lots of electricity during it and supports the grid price keeping more generators profitable, then takes the highest price and sells into it, keeping more uses reasonable. A key point here is that almost all electricity transmission connections are bi-directional and the cost of making them so is a trivial addition on uni-directional transmission.
By putting some demand (the data centre) together with some storage (the batteries) together with some generation (wind or solar plants) then at the moments that really matter each of the elements benefit. Importantly, the batteries, and especially the renewable plant get to be overbuilt for the data center because they can also satisfy other needs. The transmission gets to share costs between incoming power for the data center (when there's no sun) and outgoing power from the panels (when there's lots of sun). When there's lots of power and lots of demand, the grid becomes transmission limited. At that point the data center us using excess power from the solar panels that wouldn't otherwise fit through the transmission and so it gets cheaper energy and the reneable generator gets more money.
It's a win, win, win situation even though you never run the whole data center for a long time on just the batteries.