The switch from linear to cyclic electron flow has long been thought to rely on the migration of antenna proteins from Photosystem II to Photosystem I. Takahashi et al. report that this is not the ...