Several general anesthetics produce their sedative effect by activating endogenous sleep pathways. We propose that general anesthesia is a two-step process targeting sleep circuits at low doses, and synaptic release mechanisms across the entire brain at the higher doses required for surgery. Our hypothesis synthesizes data from a variety of model systems, some which require sleep (e.g. rodents and adult flies) and others that probably do not sleep (e.g. adult nematodes and cultured cell lines). Non-sleeping systems can be made insensitive (or hypersensitive) to some anesthetics by modifying a single pre-synaptic protein, syntaxin1A. This suggests that the synaptic release machinery, centered on the highly conserved SNARE complex, is an important target of general anesthetics in all animals. A careful consideration of SNARE architecture uncovers a potential mechanism for general anesthesia, which may be the primary target in animals that do not sleep, but a secondary target in animals that sleep.