Mitochondria are increasingly recognized as central players in the life and death of cells and especially of neurons. The energy-dependence of retinal ganglion cells (RGC) and their axons, which form the optic nerve, is singularly skewed. In fact, while mitochondria are very abundant in the initial, unmyelinated part of the axons anterior to the lamina cribrosa, their number suddenly decreases as the myelin sheath begins more posteriorly. The vascular system also presents different blood-brain barrier properties anterior and posterior to the lamina, possibly reflecting the different metabolic needs of the optic nerve head (unmyelinated) and of the retrobulbar optic nerve (myelinated). Mitochondrial biogenesis occurs within the cellular somata of RGC in the retina. It needs the coordinated interaction of nuclear and mitochondrial genomes. Mitochondria are then transported down the axons and distributed where they are needed. These locations are along the unmyelinated portion of the nerve, under the nodes of Ranvier in the retrobulbar nerve, and at the synaptic terminals. Efficient transportation of mitochondria depends on multiple factors, including their own energy production, the integrity of the cytoskeleton and its protein components (tubulin, etc.), and adequate myelination of the axons. Any dysfunction of these systems may be of pathological relevance for optic neuropathies with primary or secondary involvement of mitochondria. Leber's hereditary optic neuropathy (LHON) is the paradigm of mitochondrial optic neuropathies where a primary role for mitochondrial dysfunction is certified by maternal inheritance and association with specific mutations in the mitochondrial DNA (mtDNA). Clinical phenocopies of this pathology are represented by the wide array of optic neuropathies associated with vitamin depletion, toxic exposures, alcohol and tobacco abuse, and use of certain drugs. Moreover, the recent identification of mutations in the nuclear gene OPA1 as the causative factor in dominant optic atrophy (DOA, Kjer's type) brought the unexpected finding that this gene encodes for a mitochondrial protein, suggesting that DOA and LHON may be linked by similar pathogenesis. Polymorphisms in this very same gene may be associated with normal tension glaucoma (NTG), which might be considered a genetically determined optic neuropathy that again shows similarities with both LHON and DOA. Exciting new developments come from first examples of mitochondrial optic neuropathies in animal models that are genetically determined or are the result of ingenious engineering of mitochondrial gene expression, or from biochemical manipulations of the respiratory complexes. Even more exciting is the first successful attempt to correct the LHON-related complex I dysfunction by the allotopic nuclear expression of the recoded mitochondrial gene. There is hope that the genetic complexities, biochemical dysfunctions, and integrated anatomical-physiological cellular relationships will soon be precisely delineated and that promising therapeutic and prophylactic strategies will be proposed.