Well, sure, storage costs depend in part on how long you have to store the energy, and how often. But surely someone has computed some kind of curve about how much storage is needed for PV to displace increasingly large percentages of peaker and baseload generation, assuming no demand response or other stabilization measures like rolling blackouts?

I hope you aren't making the assumption that storage = batteries, then arguing because weeks and months of batteries would be needed, it's impossibly expensive.