Cost reduction can occur with increased demand. However there is also a flip side to this. A build it cheaper and the market will grow. This is the tick tiock model Intel followed for years. Every 18 months change the process and design to lower cost and improve performance. So an $8 billion bet that tossing out the manufacturing tools and replacing them with more and fast transistors per processor would grow the market and keep profit margins high. This was a step function, a leap every 18 months or so.
In batteries yes the cost decline with volume for the same basic product.
However there are effort to toss out one type of battery and replace it with a new type. Moving to sodium or solid state are two examples, moving to LFP (lithium Iron Phosphorus another.
So what? Essentially we are likely to see jumps in cost or performance that can spike demand.
The investment to toss out the old factory and build a new one compared to continuous improvement of one factory suggests big bets are needed to stay in the game.
And predicting the pace of change may be more like semi conductors then microwave ovens.