Demand for solutions providing power for days simply does not exist currently. Hours is good enough. This business is about providing lots of power quickly.
You are right that they typically are only used for a few minutes to deal with peaks in demand. That's because they are only needed for that long and because you can turn them on and off pretty much instantly. You don't have that choice with traditional solutions like gas plants.
Which is why despite battery cost, this type of battery is very successful at largely removing the need for having peaker plants. It also seems to be very successful at getting rid of blackouts because there were never enough of those peaker plants to begin with in Australia. Another advantage is that you can put these batteries all over the place. Many small installations make up for a lot of aggregate capacity. You can put them in buildings even; and people do.
Musk is also building factories that are producing batteries by the GWH per year per factory and soon TWH per year. He's not alone of course and basically there are quite many other companies now planning their own little GWH production facilities. There are tens of billions being invested in that stuff in the next few years. There will be hundreds or thousands of these factories in a few decades churning out many TWH of storage annually.
That battery capacity is of course mostly intended for the transport sector and not for grid storage. But if you are worried about having enough storage capacity on the grid, there will be multiple TWH of batteries driving around on roads in most countries by the end of the decade. Basically 2M cars with 50KWH batteries == 1TWH. That's a lot of distributed and mobile storage. All we need to do is plug it in.
Of course there will be more optimal solutions as well. But it's capacity we'll have none the less.
