Right now it takes a $10,000.oo rig to run the stupid demo, so you can figure it would take a console with about half that much power. In other words, at least another 3 or 4 years before it becomes practical. In the meantime the more demos like this that are available the better for the engineers trying to design the consoles.
This engine is really the brute force approach and why you don't see more demos like this. Just adding more computationally demanding lighting, depth of field, and particle effects then doing anything really tricky. What will really blow your socks off is when programmers figure out all the other trickier things they can do with that much computational power.