World Lithium Supply (2010)

But that said, the real "win" comes when we have capacitors which are mechanically storing charge.

Today you have two choices, one is a "plate" capacitor which builds up a separation of charges between two plates, limited by the dielectric constant of the material holding the plates apart, and chemical separation of charge by binding electrons to the atomic structure of two materials that can will exchange electrons given the opportunity.

Mechanical storage of charge will probably involve a nanostructure of dendrites and spheres which are filled and emptied using magnetic fields. An early example was magnetic bubble memory which the presence or lack of charge in a small bubble, moved through a "racetrack" by manipulating the fields around the bubbles could be "read" or "written". If the charge carrying capacity of the bubbles was significantly improved, one could imagine a device where each bubble stays within the dielectric limits of carrying charge, but the total population of bubbles holds significant charge. As with the chemical process an internal "resistance" is present by the mechanical limitation of pulling charge out of the device (that is the rate of chemical reaction in a battery) but unlike a chemical battery such a device would by "fillable" at the same rate as it was "drainable".

First, the calculation of the available supply is based on known, existing reserves economically retrievable at current prices. Changes in demand will change both the known reserve quantity and the price. Lithium is currently about $1/kg, so there is plenty of room to grow price to increase supply without affecting the cost of batteries. So a reasonable estimate of lithium supply has to include both demand volume and price. Reserve growth based on demand volume and price is a very tricky thing to forecast.

Second, the choice of 3.4 billion vehicles to "bring the entire world to North American levels" as a goal is an unrealistic number for many reasons. Regional economics, geography, and social density vary quite radically around the world. The built infrastructure varies and supports many different transportation modalities. It also doesn't take in to account changes in technology, infrastructure, and social organisation during such a radical transition. So a more reasonable estimate would be to simply replace all vehicles on the road with BEVs.

Last, the note could also be improved by doing sanity checks. For instance, could the world production capacity of any other energy storage mechanism, including oil, hydrogen, ethanol, etc. support 3.4 billion vehicles? It's always useful to compare what you are calculating to the alternatives.

There are currently about 1 billion cars in use in the world. Using just the reserve numbers quoted in the article it seems economically feasible to replace them all with Lithium chemistry battery electric vehicles.

First: most cars are used for urban journeys, and most urban journeys are relatively short-range (roughly 12km in the US; half that in Europe). You could have a vehicle with a 30km range and it would suffice for >95% of your travel needs. Hydrocarbon fuels, with their high energy densities, made range a secondary design consideration -- there was little downside to designing for the 99.9th percentile of use cases. But range is fundamentally more costly in battery-based vehicles, which changes the design logic. I expect both users and the industry to recognise that it's better to design for the 95th percentile of use-cases, and rent larger vehicles for the rest.

Second: infrastructure will be developed which mitigates/eliminates range concerns. Things like in-road inductive charging on motorways would make intercity travel possible even on a tiny battery.

Third: In contrast to most of the 20th century, in the 21st century, cities are becoming denser and more walkable, public transport is generally becoming better, and younger generations are less enamoured with the car than they used to be. These trends don't look to reverse anytime soon.

Fourth: Cars which require drivers spend > 95% of their lives parked. Vehicles which don't require drivers can provide taxi-like on-demand personal transport -- but as a public mode, for a fraction of the price of an actual taxi -- and they can be in essentially continuous service. Depending on the nature of the demand patterns and consumer acceptance of ride-sharing, one Robotaxi can do the job of 20-40 cars. Given the cost and convenience of such a transport system, it can be reasonably assumed that many people would forego private car ownership altogether, and rely on Robotaxis exclusively. While this wouldn't take cars off the road per se, it would certainly reduce vehicle sales dramatically (while rendering parking lots anachronistic).

I expect that these trends, over the next 30-40 years, will produce a roughly 20-fold reduction in car ownership in the developed world. This will be offset somewhat by still-rising car ownership in the developing world, but the bottom line is that the industry is likely to shrink down to a size which the earth's lithium supply can handle.