Human recessive osteopetrosis (ARO) represents a group of diseases in which, due to a defect in osteoclasts, bone resorption is prevented. The deficit could arise either from failure in osteoclast differentiation or from inability to perform resorption by mature, multinucleated, but nonfunctional cells. Historically, osteopetrosis due to both these mechanisms was found in spontaneous and artificially created mouse mutants, but the first five genes identified in human ARO (CA-II, TCIRG1, ClCN7, OSTM1, and PLEKHM1) were all involved in the effector function of mature osteoclasts, being linked to acidification of the cell/bone interface or to intracellular processing of the resorbed material. Differentiation defects in human ARO have only recently been described, following the identification of mutations in both RANKL and RANK, which define a new form of osteoclast-poor ARO, as expected from biochemical, cellular, and animal studies. The molecular dissection of ARO has prognostic and therapeutic implications. RANKL-dependent patients, in particular, represent an interesting subset which could benefit from mesenchymal cell transplant and/or administration of soluble RANKL cytokine.