PV is priced per kW. Storage is priced per kWh. You typically merge them by fixing a storage time in terms of the rated power of the collectors. This works quite nicely since the geographic insolation information is also listed in terms of hours of peak power. Consensus estimates for battery pack prices are about \$68/kWh by 2026 for EVs and BESSes. The cost per kWh of battery throughput is a more useful measure when deciding if a BESS makes sense (the \$117/MWh projected figure). As a first pass, you can take the rate spread from the utility as a target cost of throughput and take the consumption during peak hours as a target capacity and use that to scope the BESS to get a general idea of what a minimal system might look like.
The reason the price difference is small is that in most applications, BESS doesn't quite pay for itself with peak shaving, so most BESSes are quite small. Their daily throughput is much less than the daily power generated by the panels. I expect the utility rate schedules will adjust as storage costs drop so as to maintain this relationship. Utilities don't really want an inherently stable grid - they want it just unstable enough so that they can justify their monopoly based on the grid management services they provide (which is to not fix the problem because it would jeopardize their business model).
Typical storage times for BESS are measured in minutes, not hours, to maximize the specific throughput and minimize the unrecovered costs of the systems.
