Dead man walking.
It doesn't matter about the roadmaps or who's right about the Itanic future here.
While, as others have pointed out, this will likely make little difference to the current HP shops in the short / medium term, although data centre refresh cycles are key here, what it will immediately kill off for HP is any large scale new business or poaching from competitors. What's the point of choosing HP for your big server strategy if you know damned well that their Oracle roadmap has a big No Entry sign within sight on it?
Everyone of any size has *something* important (more usually lots of somethings important) on Oracle. Porting databases is hard and you really don't want to do it if there's another option. Even then you are still SOL on the Oracle applications.
As I see it, HP need to do something and they need to do something quickly. I see four options, none of which are pleasant.
1) They can give up and become another commodity x86 player. That'll hurt.
2) They can go cap in hand to IBM for Power processors. That makes sense, with two of the major players on Power, the likelyhood of the software lads doing the dirty on it benefits both HP and IBM. Also here the Power chips have a habit of turning up on time, removing one of the main Itanic gripes for HP customers.
3) They can go cap in hand to, er, Oracle for SPARC processors. It not being wintry in Hell, I can't see this happening.
4) They can resurrect PA-RISC or come up with something new and hope the software lads stay / jump on board. Dodgy.
The elephant in the room is that in HP land, they haven't yet completed the move from PA-RISC to Itanic and even on the Itanic chips a lot of what's running is in PA-RISC emulation mode. Having to perform another architecture change while the last one's still ongoing would be something of a pain for them I am sure.