> Exponential drop in the cost of solar/wind is good, but will not necessarily continue for long
Maybe not, but the predictions on this have tended to do linear extrapolation of a clearly exponential trend. Things like this:
https://pbs.twimg.com/media/DAXDEAYXgAA7up_.jpg
We won't be exponential forever, an S-curve will show up eventually but the conservative estimates are not even close.
> and to become a large % of the grid will also need to come with gigantic storage projects
This isn't really true. There are already grids with 60%+ renewables without any storage projects to speak of. Large grids already have a lot of diversity and the existing hydro is already a big buffer.
> Lots of pumped back hydro? Trillion dollar projects, and we have water crises all over, so we can't just dam up any river that it is convenient to do it either.
Most of the time those kinds of estimates are predicated on the assumption that we must produce exactly what we consume and thus storage needs grow exponentially. If you have a really really cheap source of power (right now solar) and your problem is that you can't store it, just overproducing it is a better solution. The reality is that we will do a mix of many things but all the simulations I've seen show this to be well within even our current technology and cost. Most claim the total cost of the resulting energy goes down not up.