[Hermitian909]

> but I temper my expectations with the realities of both cost and difficulty to mass produce

This is a good instinct but, FWIW, I'm think there's a good chance we're going to see scale-up costs drop by about an order of magnitude over the next 5 years via fairly straightforward use of ML to reduce on scaleup test iterations. (For readers without context: to scale up battery production you start by making a few grams of materials in a lab, then a few kilograms, then a few tons. At each step there are failures and you have to iterate on production parameters, testing can take a long time as well. Sometimes the whole effort fails). At my last job we worked on this and saw good results, but productionizing them was a ways off.