My general thinking is that waiting for any certain tech is pointless, as soon as it's released you'll want to wait for something else. Get the best of what you can now, or atleast what you can afford, and get ready to do it again in a few months or a year.
Especially for SATA, jumping on the bandwagon as soon as it's released will just make you a guinea pig for the first chipsets. And I believe SATA is supposed to have backwards compatibilty so you could use the old drives on the new controller until you get cash to get new drives. I'm also of the thinking that you should just go with SCSI, ATA is a bad design with many bandages to cover it up, they need to replace it instead of adding more kludges.
If you want a 64-bit processor get an Alpha for a UltraSparc, you can get whole systems of either for round a grand and you'll get a real, proven 64-bit OS too ( Tru64, Solaris or Linux). I wouldn't run XP 64 for anything critical for quite a long time, the 16->32 bit transition was bad for MS and I doubt the 32->64 bit one will be any easier. But in general you won't see any benefits from a 64-bit chip, unless you work with extremely large data sets.