I don't need one right now but will do at some point, but I can wait until I can purchase one. The two projects I'm working on are
1) A lightweight 2D Core + maybe a 3D core. The ADI dumb frame buffer core is just silly and huge IMHO, so replacing that is coming along nicely. After that would be nice to have enough acceleration for wayland/DirectFB in a small MIT'ed open source core, so even the BSD guys would get the benefit. With the ticket to ride core now open sourced, it kinda knocked my desire to continue a complete 3D core beyond the personal achievement desire, though implementation using the Epiphanies for the shaders (possible also portions of the texture samplers) would be good.
Coming from a games dev background with articles etc. published on graphics and shaders would be nice to be able to say I've written it from the HDL upto fancy fx personally. Kind of like, they say you should make a CPU once in your life... also would be nice to be able to tell IHV GPU vendors actually I do know how hard it is to make HW do X when I ask for something

2) Porting Sproulls FPGAtools to zynq. Xilinx tools don't work most of the time for me and the crappy partial synthesis issue (its an expensive option, when imho its whats really cool about FPGA) make me want an alternative, Torc and fpgatools show its possible, soon, maybe if someone sits down and puts the work in. I've been hacking / figuring out for a bit and I'm now starting to take little hacks into a working framework. So far the checked in code (on my fork) is mostly re-org of sproulls code (he's been unable to extend his work due to time constraints) but as time allows i'm hoping to generate zynq 7020 bitstreams.
https://github.com/DeanoC/fpgatools