A large earthquake unlocks a fault-zone via dynamic rupture while releasing part of the elastic energy stored during the interseismic stage. During an earthquake, the fault-zone weakening and associated energy dissipation occur in two temporal regions: the leading rupture front and the frictional zone at the front’s wake. Based on analyses of experimental observations, we argue here that the processes at the rupture front control the earthquake instability. First, recent experiments with frictional interfaces in brittle acrylics and rocks have explicitly demonstrated that prior to nucleation, frictional interfaces can be `overstressed’ relatively to the post-earthquake, residual stresses. Second, the experimentally observed singular stress-fields and rupture dynamics are precisely those predicted by fracture mechanics theory. Third, detailed experimental measurements of dynamic slip along a viscoelastic interface display the development of pressure and shear shock fronts that are associated with extreme strain-rates (4,000/s) at the rupture tip. Rock-mechanics experiments indicate that typical rocks fail by fragmentation to become cohesionless, pulverized material when subjected to similar shock conditions of strain-rates and stresses. Now, if these shock conditions develop near the tip of an earthquake, the fault-zone is expected to breakdown and weaken within the dynamic, fast propagating rupture front. We therefore argue that earthquake dynamics are best understood in terms of dynamic fracture mechanics and are not governed by the frictional properties of faults. In this view, rupture dynamics are driven by the release of part of the elastic energy at the rupture front, whereas the post-earthquake residual stresses and the energy dissipation are governed by fault frictional properties.