I'm not sure batteries are the answer on a grid scale. From the grid/market perspective, they're incredibly expensive energy time machines. They also come with their own environmental issues. Batteries make sense if you need to make small islanded systems, or need black start capabilities. If you just want to teleport energy/inertia around I think the answer is old school flywheels, pumped hydro, and synchronous condensers.

The real lowest hanging fruit is industrial demand-side response. I'm actively working on this from the grid side, implementing partial control room scada control of heavy industry for real-time grid balancing. Costs a few engineers some grey hairs, which is a lot cheaper than building grid scale batteries to provide the same end result.