It goes along with improved silicon processes. Each new process double the number of cells available which means that the capacity at a given price point is more or less doubling every 2 years. New processes tend to be more expensive as the yields are somewhat lower and then as it becomes reliable the prices drop inline. So you are going to see the prices of SSDs jump as new process tech comes out.
Where it gets muddied a little is that all the different manufacturers of NAND are at different stages in their process. They aren't perfectly aligned with each other and they are also doing it different. Samsung just released a new 40nm process, those capacitive transistors are huge compared to their competitors and a lot more durable. Yet they also add 3D stacking which means that a single chip has quite a lot of capacity.
The price points for SSDs more or less follow the channels available. We have had for a while 2 - 8 channel SSDs with the performance and price varying depending on how many NAND chips were used (1 per channel). As the density increases what you find is that the same channel width now has twice the capacity it did before, but it also often means the performance now moves up to the next capacity as well.
Its just how the silicon industry works. Each new silicon process is meant to bring increased transistors at the same price point, in CPUs/GPUs that translates to performance and in SSDs it translates to capacity.