White adipose tissue (WAT) is abundant in humans as a major means of storing neutral lipids whereas brown adipose tissue (BAT) is associated with an insulin-sensitive phenotype likely achieved through secreting beneficial factors and possibly by enhancing thermogenesis. Interventions that transiently cause “browning” of WAT to produce “beige” adipocytes can enhance glucose tolerance, as does implantation of BAT or beige adipocytes in obese mice. Here, we aimed to permanently convert white adipocytes into “beige”-like adipocytes by genome-editing with sgRNA-Cas9 ribonucleoprotein (RNP) complexes under conditions that avoid uncontrolled integration of vector DNA in the genome, immune responses and off-target effects. We developed a novel CRISPR-based method to deliver Cas9 protein/sgRNA complexes ex vivo to strongly enhance a beige-like phenotype with virtually 100% efficiency in mouse and human adipocytes. This was achieved by targeting the thermogenesis suppressor gene NRIP1 in preadipocyte progenitors expanded from human and mouse subcutaneous adipose samples prior to their differentiation. With our optimized RNP delivery in progenitor cells, we targeted various genomic loci spanning the coding regions of the mouse and human NRIP1 gene, and identified precise target sequences that cause maximum upregulation of beneficial secreted and thermogenic factors including UCP1 (30 to 200 fold). Residual Cas9 protein and sgRNA were rapidly degraded and that off-target editing was undetectable in the human engineered adipocytes. Implantation of either human or murine CRISPR-enhanced beige-like adipocytes into obese mice ameliorated glucose intolerance and adiposity, prevented accumulation of liver triglycerides and liver inflammation and increased energy expenditure. Extending this proof of concept to nonhuman primates, similar enhancement of “browning” by Nrip1 disruption in such cells in vitro was achieved. Application of this CRISPR technique to disrupting PDE3B also promoted the thermogenic phenotype, which was further enhanced in double Nrip1/PDE3B CRISPR disruptions. These findings show efficacy of CRISPR-based genetic modification of human and mouse adipocytes to improve metabolic homeostasis in preclinical models.