This whole situation seems rediculous to me. It sounds like Castle were wanting to produce a new, modern RISC OS computer, but RISCOS Ltd. couldn't provide them with a suitable OS. They therefore did the only thing they could do - bought a 32-bit clean version of RISC OS from Pace, along with all associated technology and IPR.
As for the suggestion that RISCOS Ltd are not interested in producing a 32-bit clean version of RISC OS, their mission statement claims that this is their primary aim. Check their web site under FAQ. Logically removing the reliance upon the aging Acorn chipset should also have been a top priority.
Remember that RISCOS Ltd. have had FIVE YEARS to deliver this OS. It's very poor that they have failed to do so. If companies are forced to go elsewhere for an OS that will work with new hardware they have little grounds to complain.
Now somebody here suggested that RISCOS Ltd. couldn't play ball and produce an OS for the Iyonix because Castle wouldn't give them info on the nVidia graphics chip. Well this too is bullshit - nVidia are notoriously secretive about their graphics chips, so it's highly likely that Castle weren't allowed to share such specs.
However the biggest part of the blame here should not rest with Castle or RISCOS Ltd. for this fiasco. Acorn are the ones who truely screwed up here. The license they issued for RISC OS ensured that the OS would split and fork, which is exactly what has happened. Had their been an insistence that Pace and RISCOS Ltd. work together to develop RISC OS I get the feeling that none of this would have happened.