Safe interim dry storage of spent nuclear fuel (SNF) must be maintained for a minimum of twenty years according to the Code of Federal Regulations. The most important variable that must be regulated by dry storage licensees in order to meet current safety standards is the temperature of the SNF. The two currently accepted models for defining the maximum allowable initial storage temperature for SNF are based on the diffusion controlled cavity growth (DCCG) failure mechanism proposed by Raj and Ashby. These models may not give conservative temperature limits. Some have suggested using a strain-based failure model to predict the maximum allowable temperatures, but we have shown that this is not applicable to SNF as long as DCCG is the assumed failure mechanism. Although the two accepted models are based on the same fundamental failure theory (DCCG), the researchers who developed the models made different assumptions, including selection of some of the most critical variables in the DCCG failure equation. These inconsistencies are discussed together with recommended modifications to the failure models based on more recent data.