New technologies that exhibit large (>10x) jumps in performance limits compared to incumbent technologies are radical technological innovations. This paper considers three cases that exhibit such jumps: the evolution of engines from the early 18th to the late 20th century (with the transition from coal-based steam engines to hydrocarbon-based internal combustion engines); the evolution of electric switches from the early to the late 20th century (with the transition from vacuum tubes to transistors); and the evolution of batteries from the mid-20th to the 21st century (with the potential transition from electrochemical batteries to quantum batteries). I show that each such transition goes along with a change in the content of the knowledge domain that maps onto the respective application area, a process that has been called redomaining.