Looks like they go after companies mainly for talent and IP. Facebook's biggest asset is their network effect and 500M eyeballs, not the software or talent, though I'm sure they have a lot of talent. Now that they are designing their own chips, after the PA Semi and Intrinsity buys, maybe silicon would fit the bill as a big aquisition for them - AMD (4.7B) or Nvidia (6.5B).
There would still be the problem that AMD always lags behind Intel in fab technology, which is why they have trouble going after the highest-performance and lowest-power segments. No matter how much cash you have on hand, there's no quick fix for that. At least with Intel CPUs, Apple products can generally match the performance of any competitor, and any price premium can be attributed to the better industrial design, integration, and software. If Apple switched to AMD processors, most of their computers would now be afflicted by a performance deficit in addition to the existing "Apple tax".
NVidia doesn't seem like a tantalizing acquisition for anybody except perhaps Intel, if they can admit that they don't know how to make a good GPU. NVidia has been mostly squeezed out of the chipset and integrated graphics markets by Intel's shady tactics and the awkwardness of AMD now owning ATI, and the discrete graphics war hasn't been going well for NVidia for quite some time, with no respite in sight. (This month, AMD will be releasing their third consecutive generation of GPUs that soundly beat the competition in price/performance comparisons at most price points. NVidia can't really do anything about it in the short term, because their chips are simply bigger for the same performance, so their profit margins are lower. They've been trying to mitigate this by focusing on HPC, but that's not catching on quickly enough, either.)
Apple could potentially transition parts of its line to AMD processors, without completely abandoning Intel.
It seems likely that Sandy Bridge will be the CPU of choice for workstations, while Bulldozer will be the throughput winner (hence server oriented), and Bobcat the performance/watt featherweight. Of course, none of this is certain until the actual products are available.
Bobcat might not have the performance to match up to Sandy Bridge, but having an AMD/ATI graphics processor could potentially make it a good choice for a souped up AppleTV product aimed a little higher. Ok, quite a bit higher. Serious 3D gaming higher. And I'm pretty sure that AMD would LOVE a design win like that!
The competition from AMD plus the the narrowing margins in CPU's is part of what's leading AMD and Intel toward system on a chip solutions, but although that's bad for the likes of nVidia, it's good for OEM's like Apple.
AMD is worth way more than their market cap, right now their market cap is depressed because of the tremendous debt. if apple bought AMD, they would probably lay off 2/3rds of the company, retain IP and pay off the debt; which is between $2-$7billion depending on how you count the cards. http://www.fool.com/investing/general/2010/07/26/amd-reshuff...
A better economy might double Nvidia's earning's, so their share price might triple.
Regarding the CPU market, I assume you mean x86 CPU market vs the ARM CPU market they are already a player in. Moving into that market might move their price a little, but increasing it tenfold? When they are entering an established, mature market against a very strong and aggressive player like Intel?