Yes, I expect that there is an orbit with unbounded velocity but I do not know how to prove it. It is indeed one of the simplest "natural" Hamiltonian systems where KAM does not apply, but on the other hand,there is no obvious way to show nonexistence of invariant curves. There might be some techniques that I am not aware of.