The nucleation rate for the precipitation of copper selenide, which has an extremely low solubility, was measured. The measured values of the kinetic parameter A and the thermodynamic constant B were compared with those calculated from theoretical considerations based on the classical nucleation theory. These calculated values were found to be physically meaningless. The nucleation work and the critical nucleus size were also found to be extremely small. It is proposed that a new conceptual model of nucleation in highly supersaturated solutions needs to be considered. A new model from literature, that proposes stable pre-nucleation clusters at high supersaturations, is presented and found to be promising. This new model would mean that the classical nucleation theory of stochastic attachment and detachment of monomers would no longer apply. The new model would also account for the very small nucleation work and the small critical nucleus sizes found in this study.