Can't see how electrical grade copper cable will be embrittled by a very fast casting of a copper alloy slug to weld two sections together. If anything, I'd expect the copper to soften by being annealed. From a Cadweld spec sheet I see that the thermit reaction is copper oxide [oxidant] and aluminum powder [fuel], with fluorspar flux, tin, vanadium and copper powder. In buried environments, a molecular bond has to be better than a mechanical crimp, which is subject to electrolyte corrosion at the connexion.
