3 ms·

> For example, the claim is that lithium ion batteries are expensive at grid scale...but the cost of such batteries continues to fall The rate at which they ha

by u320 4y ago

> For example, the claim is that lithium ion batteries are expensive at grid scale...but the cost of such batteries continues to fall The rate at which they have been falling has been diminishing for quite some time, until this year when they are starting to rise (more than inflation). The idea that past price movements will always continue indefinitely has zero basis in reality. And yes li-ion batteries are way too expensive for large scale grid usage, even if they did fall at 10% yearly we would be decades away for them to be a satisfactory solution.

Schroedingersat 4y ago
> The idea that past price movements will always continue indefinitely has zero basis in reality. This is true. Any pretend exponential curve is at best a logistic curve near the beginning or middle. But there are good reasons to think batteries as a whole are in the class of logistic curves near the middleish. There are lots of chemistries in the 500-1000Wh/kg range with various deal-breakers making them not an option. Chemical reactions go up to about 10,000Wh/kg and some of them as yet undiscovered might resemble 'battery' more than 'fuel cell' or 'magic portable nuclear thingy'. Simple few component mass produced objects made of abundant materials once every ounce of cleverness in production is wrung out tend to cost around $1-10/kg plus their energy cost. So if we set the upper asymptote at around $1-10/kWh, old technologies are around $1000 per usable kWh and LiFePO4 is around $300/kWh for final products you can actually buy. It seems reasonable to posit that $20/kWh is achievable before going too far past the inflection point. This is in line with CATL claiming $60/kWh for SIB in the late 2020s, then having costs taper off and settle around $30/kWh in 2040 as the low hanging fruit are all gone. Maybe there'll be a new discovery around then that pushes it down to $1, but it seems unlikely.

by u320 4y ago

Schroedingersat 4y ago

> The idea that past price movements will always continue indefinitely has zero basis in reality. This is true. Any pretend exponential curve is at best a logistic curve near the beginning or middle. But there are good reasons to think batteries as a whole are in the class of logistic curves near the middleish. There are lots of chemistries in the 500-1000Wh/kg range with various deal-breakers making them not an option. Chemical reactions go up to about 10,000Wh/kg and some of them as yet undiscovered might resemble 'battery' more than 'fuel cell' or 'magic portable nuclear thingy'. Simple few component mass produced objects made of abundant materials once every ounce of cleverness in production is wrung out tend to cost around $1-10/kg plus their energy cost. So if we set the upper asymptote at around $1-10/kWh, old technologies are around $1000 per usable kWh and LiFePO4 is around $300/kWh for final products you can actually buy. It seems reasonable to posit that $20/kWh is achievable before going too far past the inflection point. This is in line with CATL claiming $60/kWh for SIB in the late 2020s, then having costs taper off and settle around $30/kWh in 2040 as the low hanging fruit are all gone. Maybe there'll be a new discovery around then that pushes it down to $1, but it seems unlikely.