THAT's better than simply taking that money and investing it into the division?
I don't know, that could just be throwing good money after bad. This isn't a software division, it's not even like their server hardware division, it's chipmaking. It's kind of a go-big-or-go-home game where your competitors -- well-funded types like, say, Intel -- can easily pour many billions of dollars into next-generation fabrication processes and equipment which will readily put any half-assed investment to shame. I don't think IBM's chip business has the customer base to make "go big" profitable, or any reasonable plan to acquire new customers, so "go home" makes a lot of sense here.
Now, the wisdom / folly of gutting the rest of IBM's various divisions is left as an exercise to the reader.