Consumable electrodes are actually doing chemical work in the aluminum. It’s not just wear from use. To produce 78 grams of Aluminum directly requires 48 grams of Carbon electrodes (using made by graphitizing some tar or coke) which oxidizes to 176 grams of CO2. Aluminum is 31MJ/kg specific energy. Assuming 70% efficiency conversion from aluminum to hydrogen and 70% from hydrogen to electricity (both fairly optimistic figures), you’re talking about 535 grams of emitted CO2 for each kWh of electricity out. Ignoring the electricity input, inefficiencies and losses in making the aluminum, and assuming perfect conversion of the tar to electrodes. Worse than a good, advanced combined cycle natural gas plant (which can get as good as roughly 350 grams per kWh for record efficiency plants).
I just don’t see how this is gonna be a net win without carbonfree electrodes. It seems just about impossible here. Storing hydrogen directly looks to make way more sense.
And?
What's the benefit? We could store energy in concrete if using gravity or use water to produce the hydrogen for the fuel cell.
The benefit is you can store energy, in aluminum. It has a high energy density, you can safely store it anywhere, it requires no containment at all, etc.
Your last statement includes the answer to your own question. They are using the hydrogen, from water reacting with aluminum, to power a fuel cell.