New drug discovery (NDD) is a fascinating discipline encompassing different facets of medicine, pharmacology, biotechnology and chemistry. NDD is very often restricted by efficacy or safety problems of the new clinical candidate in human patients. Drug regulatory authorities have provided various guidelines for advancement of safe new chemical entities (NCEs) in clinical trials which must be strictly followed. In spite of this, various drugs have failed in clinical trials or withdrawn from market because of human safety issues related to cardiotoxicity, hepatotoxicity, neurotoxicity and teratogenicity. The failure of safety prediction was pointed to species specificity issues, lack of mechanistic toxicity data and inadequate clinical trials. These drugs not only affect human health but also cause loss of resources and time. The species specificity issues are partially addressed by use of primary human cells but their availability is very limited. Human embryonic stem cells (hESCs) and induced pluripotent stem cells (hiPSCs) offer sources for generation of an unlimited number of human somatic cells. The emergence of mechanistic models for toxicity testing with transcriptomics, proteomics along with toxicokinetics readouts based on hESCs and hiPSCs is paving the way to design new human relevant testing strategies. Introduction of these models at the timeframe of lead selection and optimization in parallel with in vitro pharmacokinetic studies will significantly reduce compound attrition rate by selection of safer lead molecules. We focused on upcoming hESCs and hiPSCs based toxicity testing models and their future role to address safety gaps of present drug discovery and development.