I have a few views on this.
The first is that anything is possible. The requirements (usually time and money) are the biggest barriers. This would take a huge amount of time and at this point not enough is known about certain internal workings that would get this working.
More knowledge of the inside of OS 9 is needed and we are gaining this, it just takes time and a lot of it. Not only that, but man hours are expensive monetarily.
Right now it's easier to use the G4's to gain this knowledge as all of them are of the same basic hardware (this is a very extreme generalization). Once the all G4 hardware is supported then it's possible to move on to the G5.
As far as the G5 though, the earlier once are more likely to get working. The later models that ditched the PCI bus will have some more difficulty.
So here is where the G5 sits and will sit for some time. The trampoline (second stage boot loader that sets up a lot of hardware) is the first major road block. Knowledge of it's internal workings are theoretical at best right now. We are working at it as it will absolutely be needed to even get things like the mini and other partially working systems (Xserve among others). That will be the first step to getting it all working on a G5. From there, I actually don't think it will be extremely difficult to get the rest of the system working, but there would still be significant work on the nanokernel and 68K emulator. By the time we get to even attempting G5, I do believe the knowledge will be there and will just be time needed to write correct drivers for the hardware.
Now here is the reason for my thoughts above.
As of right now it is possible to boot a G5 up to the point of it starting the nanokernel at which point it fails so early that getting debug output is basically not an option. To even get some output will require the hardware to be setup and running. To get this happening about 80% of the hardware has to be disabled or removed from the device tree (mostly removed). Very important hardware had to be disabled such as the memory controller (not to be confused with the MMU on the CPU), PCI buss, power management, all sensors, USB, FW. Video had to be greatly modified, and almost all interrupt controllers. The interrupt controllers actually don't seem to hold it, but how the hardware is configured in the G5 certain hardware beyond some interrupt controllers would only disable if the interrupt controller is disabled. OS 9 recognizes some hardware incorrectly so it really messes with stuff. It's also very dangerous if you don't want a large paper weight. When I worked on the possibility of OS 9 on the G5, my only G5 at the time almost became a brick. It was extremely hard to get it to boot as it was not possible to get to open firmware or reset the nvram from the keyboard. So is it possible to boot? Yes. Is it safe? No. Can you boot far enough to do anything? Not even close. It can't even boot far enough to finish booting.
One hardware issue is the SATA ports/bus. On the G5 only AHCI is supported and not IDE. So essentially proper full SATA driver and support will need to be added to the system. This can be done now by writing an system extension/driver to accomplish this, but that has not happened. It hasn't been done because it was far easier to have IDE busses emulate SCSI for boot and such. Once SATA really came around, OS 9 was pretty much irrelevant to the hardware manufactures. It had no value to anyone at the time to create these extensions.
So in theory, this could be worked on and developed using the G4 and non-mac SATA cards. Again think time.
Take all this with a grain of salt though as a lot of it is based on my knowledge and opinion.
If I get some time I will post the OF commands required to boot OS 9 on the G5 so that you can have your own giant paperweight. Though I do believe I owe someone instructions on setting up a late model iBook G4. That would come first.
Seriously though, the G5 doesn't do much more that crash and burn.
