Thanks for the PCB shots, some minor analysis (I am not much of an electrical engineer in this regard). I also took the back side image, scaled it and flipped it to overlay/sit beside and compare.
The main two chips there
Winbond
Looks like 25Q160VSIG but I can not find anything right now on that,
However if it is 25Q16BVSIG then W25Q16BVSSIG is the chip in question. I can see some other things being off but that really does not look like a B to me.
Still
http://www.digikey.co.uk/product-detail/en/W25Q16BVSSIG/W25Q16BVSSIG-ND/2208449 , prices say "call" but similarly potent chips from the same company are in the sub £1 range.
https://www.winbond.com/NR/rdonlyres/7EB3B29C-1B35-421C-AA24-F430B51C776A/0/W25Q16BV.pdf
A 16Mbit (2 megabyte) flash chip if the B is correct, either way a single very low capacity flash chip. Most likely programming for the other chip you see, I have not traced down everything (and if it is three layer or more then I will not be able to) but as near as I can tell there is nothing going to the 3ds slot. With this being the case any "10 game, only write once" limit is pretty clearly an artificial hobbling (even with what I am about to cover), it might be a technical limit without in field programming but we have done that before (see passme2 stuff).
Edit
25Q16CVSIG could be it, though it still does not look like C to me (more than a B but hey)
http://www.winbond.com/NR/rdonlyres/63CFB9A9-816D-4DE0-85EE-FF08F9244193/0/W25Q16CV.pdf for the datasheet. Still a 16 megabit flash chip though.
For the other I have no idea. ASIX, proflex and the model number bring back very little. Clearly it is the brains of the operation but I have nothing.
ASX093F6760 would appear to be the numbers, however that 9 might be an S, that first 6 could be G and I am not sure about the last one (could be a 0, a C or a G, or possibly something else). Not knowing this somewhat limits what I can do, especially as far as doing any kind of bill of materials/cost analysis. 88 pin QFP though and over half the pins used.
The microSD appears to be being used in SD mode (I assume SDHC does not change anything of note here) which is better than some of the things I see using SPI mode (pin 1 of the microSD, which is unused in SPI mode and a data pin in SD mode, would appear to have a via underneath that then pops out just left of X1 in top side image, which then feeds directly into the ASIX chip).
http://www.planetmobile.it/jumpjack/adattatore/pinout-tflash.jpg
No idea what SAM12 is as far as make or model, it is pretty clearly part of clock generation and voltage regulation type stuff though, to that end nothing major, though it is tied to several things.
Looking at the back side image I am going to say the PCB is possibly not a simple double sided affair (probably no more than three layers but still not double sided) -- some what is visible of lower layers on the bottom side shot has no corresponding part on the top side. I can see a few not working with my scaling (maybe some differences in angles of the shots that I have not quite accounted for) but my via holes line up well enough on both sides. Also I see what might be the stink of an autorouter, or might just be someone not used to layout trying hard in reasonably high density world.
I don't have a good 3ds cartridge slot pinout but knowing it shares a common save method with the DS I can use
http://problemkaputt.de/gbatek.htm#auxdsgamecardslot to guess a few things. I could start filling in what various pins do but without knowing more of the ASIX stuff it is not of that much use.