While humans and robots can both play the notes on a score, human expressivity—the micro-fluctuations in rhythm and note duration that impart a performer’s personality on the music, is extremely difficult for robots to mimic. Instead of forcing the robotic performers to play more like humans, Microbursts explores the ways in which robots can most comfortably express themselves. In Microbursts, AMI, CARI, and TAPI perform textures influenced by granular synthesis techniques. The speed and precision of these electromechanical instruments is harnessed to produce sonic grains on the level of microsound, whose amplitude and position in time can be stochastically determined. Modulating these parameters produces shifts between cloud-like textures and precise unison rhythms. While a human may determine the high-level control of the piece, AMI, CARI, and TAPI are free to impart their own (admittedly “mechanical”) performance style on the music.