Summary Future cost reductions in renewable generation and storage technologies hold promise for dramatic changes in the design, cost and carbon emissions from microgrids. Microgrid design depends nonlinearly on technology costs and this article analyzes this non‐linear relationship by developing an analogy between phase transitions in physical systems and microgrid design. Technology costs are akin to state variables that influence the “phase” of the system, described by the presence and activity of different energy technologies in an equilibrium (least‐cost) microgrid design. The phase transition concept is explored by modeling the least‐cost design and operation of standalone microgrids in a broad space of photovoltaic and lithium‐ion battery prices. Results indicate clearly defined phases— boundaries with sudden jumps in incorporation of solar, batteries or diesel into the system and discontinuities in the derivative of system cost at these boundaries (1st order phase transitions). For example, for a residential system in Rochester, photovoltaic prices fell such that the optimal share of photovoltaic capacity jumped from 0% in 2011 to 32% in 2012. Looking to the future, we find combinations of solar and battery prices needed for microgrid parity (levelized cost of electricity = US$0.1/kWh). Comparing these future price requirements with historical trajectories suggests attaining grid parity will be challenging. The ideas developed here, with modification, could also be applied to macrogrids.