In theory, if Wayland succeeds and I hope it does, it will be much better in some aspects. The architecture is described here:
In my humble opinion, the main point of Wayland is the following one: right now, a particular application can use OpenGL to draw graphics inside a window if it wants to, and OpenGL can be used by X compositors to create the desktop effects we're used to see in KDE4 or Compiz. However, when a normal application with an X interface wants to draw something on screen, it uses the X protocol to do that drawing. The X protocol is very basic and allows drawing simple things like rectagles, lines, etc. So when you want to update a button to look pressed or shaded, or anything, your CPU has to do the drawing. The rendering of your screen happens almost completely in your CPU.
While the X server has worked very well for years and we have to thank the different projects and people for providing us such a central piece of software during all this time, it may be a good moment to create something new, that takes into account how graphics cards work today and the power they all have, and shift the rendering away from the CPU and more into the GPU. Wayland may be able to fill that gap (refer to the "Wayland Rendering" section in the document above). The drawback is that it's a new project, it will take a lot of time to stabilize properly, and it will have to provide backwards-compatible software to slowly change how things are done today (refer to the "X as a Wayland client" section). A lot
of time will pass until we see an almost completely Wayland desktop.