I'd have thought it'd drop to at most $.50/GB by now. It's still at about $1/GB. (Apple prices: 128 = $100 baseline, 256 = $300, 512 = $600, 768 = $1000) Is it really that hard to make it? Prices have dropped, but only slowly and slightly. It'll be a great day when you can interchange the Flash with the HDD and not have to pay that much extra if anything. Whenever that is.
Am I rambling?
You are, but I sympathize with your frustration.
NAND has seen tremendous improvement in size and speed since they were first used in PC devices, but just like magnetics before it, there was a long bloodletting period where not everybody was on board. I bought a Vertex 1 some time after launch for almost $400 - now the same device with a newer chipset is closer to $100.
We're seeing more frequent improvement now that the technology is better understood and virtually every storage company in the business is using it. The market pressures of companies like Apple have forced down prices from their previous stratosphere, and now we even have 960GB drives for well under a thousand dollars.
NAND's current problem right now is scale. Since it's still silicon, it's linked to Intel and everyone else seeking a fiscally feasible way to go from 40nm down to 5nm. Die shrinks are the fastest route to more capacity, but they reduce speed and cost billions in research and retooling. Science itself also is having trouble keeping up with the pace needed - at some point all we can do is add more chips to get more storage.
The pressure has been so severe that now there are three classes of NAND just to store more data per dollar. You can double or triple the amount stored per cell...but you pay an equal price in lifespan and speed for it.
Magnetic storage has had some pretty significant advantages over time because the material hasn't changed (much). Instead, platter density has gone up thanks to a) awesome science and b) new organizational techniques. We've pretty well past what was supposed to be the theoretical limit on storage per inch. (If NAND research ever fails to make a breakthrough I think it's pretty fair to expect a surge in interest in MRAM instead.)