Assume the wedge is like a cone with curved end. At first, there is only one normal force between the tip and the material below it. You can assume the surface of the material as a thin film of very large surface tension. This penetrating force causes the film to deform and the cone goes slightly in the film (the film doesn't not tear). Now because of the curved tip of the cone, the tilted edge acts as a constraint, converting some of the normal force to side-ward force. As a result, we have the material to compress to sides, this means there must exist some where the material is stretching(, so that it's compressing somewhere). That point is the tip. Since the material can't tolerate too much stretch, the bonds break down and the material separates below the tip, this makes more space for the wedge to go inside, and the same process repeats. But this doesn't continue forever. That's because friction force is acting and because any material is at least poorly elastic. But why doesn't the friction stop the wedge at the start? The answer is that the magnitude of friction force depends linearly (classical view) on the normal-ward (I have to say it like this :)) force on the surface. The component of the penetrating force doesn't change much, but the materials response to stretch, which is compression, tries to get back the material to it's initial state, that is to squeeze the wedge, that is to produce a normal force that helps friction grow, and an overall upward force that want to shoot the wedge upward.
I tried to include any details. If some part is ambiguous, tell me.