I also thought it might be helpful to give a market-oriented historical perspective about the microSD card, for those who are curious about how this whole thing works.
Flash prices are volatile because the fabrication facilities ("fabs") that make the chips run at a constant capacity. You can't simply turn them off when demand goes away; furthermore, they are optimized to be profitable when running at near peak capacity. A fab running at half-full typically loses money due to the fixed overhead of running the equipment -- for example, the vacuum chambers must be extremely pristine, so even if a wafer isn't in the machine at the time the vacuum is always being pumped; allowing ambient, dirty air into the chamber would be disastrous. Furthermore, the cost of the blank silicon wafer itself is only about 10% the cost of a chip -- the other 90% is in the consumables, depreciation, and fixed costs of operation. Thus, when sales are slow, a fab will still make Flash memory even if there's no demand, because you lose less money selling chips at a loss than selling no chips at all. This causes a glut and drives prices down. On the other hand, if demand outstrips supply, prices quickly shoot up because it's virtually impossible to quickly increase a fab's capacity. Because the equation has fixed supply and variable demand, even small percentage differences between supply and demand tend to integrate over time and can lead to a massive market shortfall or glut. The situation is analogous to the supply and demand of oil. However, in semiconductors, pricing anomalies are compounded by Moore's law -- every time fabs shrink their process you get more memory for the same price as before, but not all vendors switch over at the same time.
The "spec" size for a chumby One firmware is less than 512 MB. Therefore, as designed, we should be able to fit into any microSD card that is 512MB or larger. Back when the chumby One launched, 512 MB memory cards were being phased out; they had shrunk the process to the point where it was no longer economical to make 512 MB cards, so 1 GB was the sweet spot. Because 1 GB was the sweet spot, a lot of large companies had contracted for large amounts of 1 GB flash. As the holiday season neared, excess consumer demand outstripped supply and drove 1 GB prices above that of 2 GB prices by about 10-20%. Thus, last holiday season, it was cheaper to ship chumby Ones with 2 GB memory cards than pay market price for 1 GB cards. Now that the holiday season is over, 1 GB card demand has dropped and prices have dropped; furthermore, my intelligence reports that a major fab has just shrunk their process, which may drive that cost even lower if demand doesn't pick up too quickly.
As I look forward to the conclusion of 2010, it's possible 2 GB cards will cross-over the 1 GB cards again; also, it's extremely unlikely that we would ever go to a 512MB size. By 2011 it's possible that 4 GB cards will cross-over as well. However, one thing is sure: this year will be very volatile for memory pricing -- all indicators show that the consumer economy is recovering and the semiconductor industry is signaling that it won't keep up with demand this year. It's currently May and certain markets are already starting show signs of a buying frenzy in preparation for the fourth quarter. Good for your 401(k), bad for chumby -- the rise in market price for raw goods will come out of our operating income. Hopefully that will be offset by any net volume increase of sales, assuming we can even get the material to build units to offer for sale. Unfortunately, we're a runt at a trough that feeds the likes of Nokia, HTC, Apple...
Due to the massive price volatility in the market, many larger consumers of Flash will negotiate fixed-price contracts. I've seen contract pricing go higher than 2x over spot pricing, so it can be a very bad deal at times (I've also seen it go the other way as well...). This is also partially why, for example, Apple will charge you 2-5x over memory cost for offering you an iPod at a higher capacity (don't even get me started on the iPad price curve...), even though technologically speaking the change is as simple as swapping out one memory chip for another. That, and the fact that it's an easy way to increase profits by segmenting the market along a simple and easy to understand consumer metric: most consumers agree with the idea of paying more for more memory because it allows them to store more songs. However, on a chumby, more memory gets you actually nothing because we have no UI mechanism for you to copy MP3s onto your device. Instead, the idea is that you use Pandora or Shoutcast to access unlimited music from the cloud. Of course, if on a future product we introduced the notion of UI-visible storage partition, we would be obligated to ship only one size of memory card, no larger and no smaller.
It might also be worth noting that the fact that MP3s show up on internal storage at all in a chumby One UI would normally be considered a bug by most companies since it is an unsupported feature -- especially since it can cause trouble down the road if we do an update that wipes the storage partition (we'd like to avoid doing that if possible but technically it's an option because the storage partition is an unsupported feature, allocated primarily for temporary files and developer use), or it can cause consumers who try to access it to be surprised or dismayed when they find that the size is actually variable, which admittedly, is a strange concept in the CE market place and requires a lot of explanation to understand.
7BAA 2E53 01C1 DCFF 497B E7F0 9699 A303 78F0 D9B9