Aggressive predictions have us generating ~6-10TWh of batteries by 2030 meaning we’re going to still need about another 3-6 years to actually satisfy demand (ignoring complexity of hooking up the batteries). On top of that, the batteries require rare earth metals that companies are gearing up to satisfy by strip mining the ocean floor for those polymetallic nodules, operations which have a very real risk of completely destroying deep ocean life. It seems to me like it’s slow and ecologically potentially more destructive than even global warming. Is it really wise to be betting on batteries at this scale vs tried and true nuclear fission which doesn’t carry any of these risks?
What people forget is batteries are a manufactured good, which follows Wright's Law. Manufactured goods (like energy storage, TVs, lightbulbs) obey different economic principles to scarce goods (like land, services, or goods with scarce inputs), and they have effectively unlimited supply. The supply is strictly set by demand.
Aggressive predictions of ~6-10TWh/year of batteries in 2030 are more predictions of demand, not so much predictions of supply. If market demand in 2030 is 30TWh/year, then that's what the market will produce. But don't blame manufacturers for the fact that demand in 2030 will only be 6-10TWh/year! And don't confuse this for a sector's inability to increase supply!
The response when seeing a "6-10TWh/year" prediction should be "how can we incentivize demand so that this number is 30TWh/year instead".
Lithium ion is preferred for vehicles because it's lighter, but again we are talking about stationary storage, so the extra weight of sodium ion isn't a problem.
The technology is solved, and the materials needed to make it abundant. It's all about demand. If the demand is there, the industrial capacity will follow. But right now, the market is only demanding about 3TWh/year of storage, and so that's how much industry is producing.
It takes a lot of time for new battery technologies to scale and disrupt existing ones and entrenched players have an incentive to continue competing. Sodium ion, iron air etc might replace lithium ion on the 30 year time scale but lithium ion will continue to drive down costs and up its capacity to try to compete and it has significantly more revenue to fund this by being the only player in the market. So it’s not clear when alternative batteries will start to replace lithium ion, but at scale it’s unlikely to be a quick process. And please don’t pretend like it’s all a demand side problem. It takes time to build out new factories from manufacturing all the equipment needed to acquiring and training employees. There’s plenty of demand for cheap batteries and the ability to manufacture simply isn’t there either and it’s being brought online. Oh and that capacity being added? It’s all lithium ion and requires a long pay off for that investment. Lithium ion is going to be potential a significant ecological debt worse than fossil fuels if the ocean floor strip mining gets going.
And it is all a demand side problem. If the world wanted to buy 10 or 20 TWh a year at current market prices, that's how much would be produced. But the world doesn't want to do that and hence that much isn't produced. This is Econ 101 for goods with non scarce inputs. It doesn't take ten years to scale up production for commodity goods.
As for scarcity, inputs to lithium ion ARE scarce which negates your entire model. Pretending they aren't is where you're making a mistake. Lithium, cobalt & nickle are relatively scarce and the mines for that have to scale up to meet demand as well. You've also got a workforce to train to do the work which takes time & is also input-constrained. That's why there's massive NEW lithium mines being opened in the US & elsewhere to extract existing reserves to meet the growth in lithium ion batteries. If the world thought that sodium ion or ion air was an immediate future, you wouldn't see these massive large-scale investments into lithium. Lilthium-ion batteries is going to be a large and growing market for decades which brings me back to the strip mining of the ocean floor that's coming to support that.
Whright's law by the way isn't also an inevitable effect that goes on forever. At some point your exponential plateau's and you no longer see such exponential decrease in pricing. That's why processors aren't getting cheaper and compute isn't scaling up quite in the same way as in the early days. There's only so efficient you can make something.