The thermal runaway of Li-ion batteries presents significant hazards to people, the environment and other assets, both physical and reputational. While rare, these events have substantial consequences in automotive, domestic and commercial applications. Hence, understanding the probability of failure and the variation in failure characteristics e.g. temperature, occurrence of fire or gas composition etc., is essential to quantify risk accurately. Many studies investigate such failure characteristics; however, they do not go to the lengths to undertake enough replicates to be statically significant to quantify uncertainty. As such, we are paving the way in experimental and computational uncertainty analysis of Li-ion battery failure by utilising our unique research facility. Through these methods, we can, for example, predict the probability of EUCAR Hazard failure level for a cell under different abuse temperatures; and the probability of thermal runaway propagation under various heat transfer conditions – showing how cell uncertainty is crucial to predicting the safe design of thermal management systems; and the probability of fire occurrence and its variation in heat flux. This presentation will discuss these findings and their importance in the risk assessment process for the design of safer batteries. Overall, the audience will leave with a greater understating of the significance of uncertainty in the context of Li-ion battery thermal runaway hazards from cell to grid level.