Li-ion cells suffer from significant safety and performance problems due to overheating and thermal runaway. Effective thermal management can lead to increased energy conversion efficiency and energy storage density. Critical needs towards these goals include the capability to predict thermal behavior in extreme conditions and determine thermal management requirements to prevent thermal runaway. This paper presents an experimentally validated theoretical model to predict the temperature distribution in a cell in response to nonlinear heat generation rate that is known to occur during thermal runaway. This problem is solved by linearization of the nonlinear term over successive time intervals. Experimental measurements carried out on a thermal test cell in conditions similar to thermal runaway show good agreement with the theoretical model. Experimental measurements and model predictions indicate strong dependence of the fate of the cell on its reaction kinetics, thermal properties, and ambient conditions. Specifically, a sudden change in thermal runaway behavior is predicted once the ambient temperature crosses a certain threshold, consistent with past experimental observations. The impact of increasing cell thermal conductivity on improved thermal runaway performance is quantified. Results presented here provide a fundamental understanding of thermal runaway, and may lead to improved performance and safety of Li-ion-based energy conversion and storage systems.