Reliability aware thermal design

Embodiments are disclosed that relate to implementing semiconductor device cooling systems that leverage awareness of regional voltage and temperature reliability risk considerations. For example, one disclosed embodiment provides a method of implementing a cooling system configured to cool an integrated circuit. The method involves first determining a heat dissipation factor that would reduce each region of the integrated circuit to a reduced temperature in order to maintain an overall failure rate. An analysis is then performed, using an insight about the relative reliability risk of elevated voltage and temperatures, to identify a region of the integrated circuit whose temperature can be permitted to rise without exceeding the overall failure rate, thereby permitting implementation of a cooling system with a reduced heat dissipation factor.