" Annual vehicle sales are 90 million, give or take. Assume 50kWh per battery, and around 850 grams of lithium carbonate per kWh. No recycling, no tech improvement, full market penetration, no savings from autonomous vehicles or the like and it is 3.8 million tonnes of lithium carbonate equivalent - market price assuming today’s lithium price of $60k works out to $230bn, or the dollar equivalent of 23 days of global oil consumption. Assume more normal lithium pricing of $20,000 and you get 8 days of current run rate oil consumption for a year’s worth of lithium vehicles at 100% market penetration. This is existential for commodity trading houses longer term: they can sell more lithium, copper and nickel but the core businesses they are in are big, lucrative and going to shrink and there is no certainty that lithium is not going to disappear into vertically integrated supply chains that do not need freewheeling intermediaries or “financialized capital”. [...] This math for lithium stands before we consider really disruptive things such as: maybe lithium isn’t it for grid storage, or even autos? This is the latest from the Sadoway lab - if we are going to use primarily sulphur and aluminium for batteries they will be laughably cheap. A tried and tested way to make money in materials science is to “do it better, with more available materials and less energy” - this is all three if commercialized. Over the last few years cobalt demand estimates have been crushed by developments in cathode chemistry due to cost and performance improvements in simpler chemistries - I am sceptical that this is the last time that today’s “unobtainium” becomes tomorrows chopped liver. Maybe new nuclear works, even if at low levels of total energy provided simply for stability and security reasons? Perhaps we can do a lot more pumped hydro than we thought? All the while the solar wafers get thinner and more efficient and use less materials…. the preponderance for everyone now calling for longer term structural energy inflation when papers like this are coming from Oxford which take account of these dynamics seems deeply unwise if you are doing anything but playing quarterly revisions. "