For starters the Astenosphere is not nearly molten let alone partially molten. It's just a wee bit more ductile than the environment like the crust and the upper mantle. The lower mantle is even more plastic again.
As far as I know, there is no theoretical limit on the height of a mountain. Those things get a bit too complicated to model mathematically. But most factors are mentioned. Buoyancy is probably prevailing. Try to think of Earth as being fluid like water. Due to rotation and gravity forces it will attain a
geodic shape, now, throw an ice cube on it and there is your mountain, the emersed part of the ice. On million years time scale Earth behaves like fluid. As the crust is much less dense than the mantle it is assumed that, in equilibrium, the dept of the crust is reflected on it's elevation, like the floating ice cube. Most of the ice is submerged. In analogy, oceanic crust is assumed to be very thin. Continental crust is much thicker to have the land float higher like the ice cube.
http://mediatheek.thinkquest.nl/~ll125/en/crust.htm
Now, if two tectonic plates collide, one will subduct, the other will pass overhead but the local thickness of the (light) double crusts increases significantly, playing buoyant ice cube on the (heavy) ductile mantle, consequently the mountain range is build. And mountains literally have deep roots of crust material (presumably).
What would limit that mountain buiklding process? Outflowing of the mountain like ice sheets do? Erosion? It's all speculation.