But with several math errors, which make him off by a decade. He multiplies by 1.59 annually and gets a 25X increase in total battery in seven years. That actually takes eight years, but whatever. Then he says that's eight doublings, but eight doublings is a 256X increase. That would take nine years at 100% annually (first year to go from 1 to 2, then 2^8=256). But we only have a 59% annual increase so getting to 256X takes about 13 years.
He also seems to be off by one in the cost reduction. At 25% per doubling it takes nine doublings to get to 10% of the current price. So add another year or two to get to $8.
It's still interesting that we could get to $8/kWh by 2040 or so, especially since it seems physically plausible that sodium batteries could get that cheap, and that we could build several days of grid storage using them. And by 2030 we still get a cost drop of almost two-thirds, down to $28/kWh if we accept his claim of $80/kWh in 2023.
