Where does solar flare energy come from? More specifically, assuming that the ultimate source of flare energy is mechanical energy in the convection zone, how is this translated into energy dissipated or stored in the corona? This question appears to have been given relatively little thought, as attention has been focussed predominantly on mechanisms for the rapid dissipation of coronal magnetic energy by way of MHD instabilities and plasma micro instabilities. We consider three types of flare theory: the steady state photospheric dynamo model in which flare power represents coronal dissipation of currents generated simultaneously by sub-photospheric flows; the magnetic energy storage model where sub-photospheric flows again induce coronal currents but which in this case are built up over a longer period before being released suddenly; and emerging flux models, in which new magnetic flux rising to the photosphere already contains free energy, and does not require subsequent stressing by photospheric motions. We conclude that photospheric dynamos can power only very minor flares; that coronal energy storage can in principle meet the requirements of a major flare, although perhaps not the very largest flares, but that difficulties in coupling efficiently to the energy source may limit this mechanism to moderate sized flares; and that emerging magnetic flux tubes, generated in the solar interior, can carry sufficient free energy to power even the largest flares ever observed.