I guess, because this is a statistical process - you can draw analogy with radioactive decay - nuclei of say Uranium-238 are generally not stable, but only some of them disitegrate at a specific moment. The reason for this is that even if the nuclei is unstable - product of it's decay are lower in energy, then the nuclei itself - then it still requires some energy for it to decay. The same is true for molecules.
It is like as if there was snow on a mountain - it want to move to the Earth's surface, because it is a more favorable position energetically, but for this to happen we first should melt it

.