If you want to acquire an entire company, and a company like IBM, you will have to pay a very high premium over the actual cap
This. When talking about Apple's cash hoard, people seem to treat it like a shopping list: "oh, I'll have this for $xBn, that for $yBn and still have $zBn for rainy days".
Actual corporate acquisitions are much more complicated. Not only is it expected that you pay a premium, but you also need to do your own evaluation of what the business is worth and what the acquisition could do to increase that value.
It makes no sense to, as some have suggested, buy IBM, take what you want and resell the rest for parts. This is a $100Bn enterprise, not a used car. The company pieces may not be viable as separate entities, or might be worth much less (making the price for the bit you want even higher).
Similarly, you just owning some part of the business might actually negatively influence its reputation. Apple is a prime candidate for that. I can't count the number of enterprise IT people I've encountered here in Germany who regard Apple with absolute disdain. They moan about how they only make products for "hipsters" and make pretty boxes full of crap. They love a bit of Microsoft and even, strangely, Google (Google have done a fantastic job of fooling the world that they support open-source software, and to a lot of IT people here that actually means something as far as reputation goes). If Apple bought IBM, or even just a part of IBM, I could see a lot of those people look for alternative providers (if there are any; I don't know the enterprise market).
A partnership would be much more palatable to these people, since they know and trust IBM as an independent entity to Apple. I'm not saying this venture will be a runaway success, but it's better than an outright acquisition and better than Apple continuing to go it alone.