Database software can be tailored for specific application domains and their required functionality, for specific hardware and its characteristics, or for other purposes. This brief paper collects issues and techniques required or desirable for making a server-class database management system energy-efficient. The opportunities go far beyond cost functions and general performance improvements. Topics include additional layers in the memory hierarchy, I/O optimizations, data format, scheduling, adaptive query plan execution, and self management in novel ways. This paper represents a challenge rather than a solution. Promising approaches and techniques are indicated throughout, but the challenge is to provide tools for building and tailoring experimental database server software that enables research into energy-efficient database management.