We laboured with this exact problem ourselves... in the end we haven't switched.... our existing domain is 12 years old with lots of long-term SEO work done on it,, the website has 4,000 pages and it just seemed like too much risk..
However if I'd considered this 5 years ago, I think I would have teken the plunge.
I also believe Matt Cutt's statement regarding KW-rich domains (funnily enough I voted for him to answer that question) is probably a little disingenuous - I think he's trying to put SEO-ers off the scent. Certainly SERPs don't seem to support what he's saying. At the end of the day, in the search for relevance, Google can no ignore the actual domain - crappy site it may be but 99% of the time it is relevant.
The other option is to develop your KW-rich name (or a bunch of them - why not?) and SEO them seperately and either generate enquiries through them or point them to your main site..
