Not all of the WCG subprojects support PPC, however; Flu, Childhood Cancer, Rice, Conquer Cancer, Dengue and AIDS support PPC, while Muscular Dystrophy, Clean Energy and HPF2 are Intel only. Other projects that I have run in the past are Einstein, Rosetta, SIMAP and Lattice.
I have an Einstein unit crunching now, and I got one or two PPC SIMAP units this time around.
With WCG, what I was implying with just a light dusting of snark was that IBM is one of the PPC standard developers. If I remember correctly, the G3 architecture was IBM. G4 switched to Motorola. Motorola spun off another group which became Freescale (I think technically "Freescale Semiconductor Inc." then just "Freescale Inc."

. Freescale still makes G4 chipsets.
* IBM then made the G5 and Power5, then Power6, and now Power7. If IBM is a major investor, projects will have PPC support. That would be like AMD or Via supporting PPC design and eschewing x86(_64) design.
Anyway, back to what you said. If you look at the projects, they make some sense. Clean Energy was a short term project. The HPF and MD projects are cyclical, and the nature of the computation can change between cycles. PPC architecture has two things to remember. First, it is not known for strong floating point computation. The desktop models (G3/4/5) are stronger on integer than floating point, since most desktop work (business, web, etc.) is integer. Low power chipsets like Via also sacrifice floating point, as it is a larger and generally hotter (as in more heat, that's literal) section of the silicon field. Second, most programming talent is with x86 design. PPC is vector based. The main thing it means is it is designed to distribute short instructions across all elements of a data set (SIMD: Single Instruction, Multiple Data). The pipeline is short with a quick release. x86 design is a longer pipeline, where you take data elements, hold on to them, and really beat the tar out of them, doing a lot of processing before release. (Geeky tangent: vector design is the same as video cards. This is why Macs were originally so good for image editing. Image transform functions are essentially vector.)
So... For the short term projects, I can see only doing x86. It is the largest and most powerful platform now. The PPC macs are fewer in number, less powerful unless talent is applied to do the design right, and dropping due to age and atrophy. (Remember, the number of Macs really didn't grow until after the Intel switch.) For long term projects, yes, I expect IBM to help with PPC design. For short term projects, I can understand why the return would be too small for the investment.