We present an evaluation of some novel hardware implementations of a stack. All designs are asynchronous, fast, and energy efficient, while occupying modest area. We implemented a hybrid of two stack designs that can contain 42 data items with a family of GasP circuits. Measurements from the actual chip show that the chip functions correctly at speeds of up to 2.7 GHz in a 180 nm TSMC process at 2V. The energy consumption per stack operation depends on the number of data movements in the stack, which grows very slowly with the number of data items in the stack. We present a simple technique to measure separately the dynamic and static energy consumption of the complete chip as well as individual data movements in the stack. The average dynamic energy per move in the stack varies between 6pJ and 8pJ depending on the type of move.