Battery power capability has long been a challenge. All common methods to determine power capability, including the hybrid pulse power characterization (HPPC) test, are estimation methods where a model is fit to measured data and power capability is calculated rather than measured. Such estimation methods are inherently error prone due to the complex nonlinear characteristics of batteries. For LiFePO4 batteries for example, power capability may be underestimated by 47%. To address this shortcoming, a generalized method is proposed to measure (rather than estimate) battery power capability. The measurement is made by iteratively applying a series of power pulses until the battery reaches one of its operating limits. Any length of power pulse is supported (e.g., 10 seconds), along with voltage, current, and temperature-limited power capability. Each successive power pulse is calculated using a physics-based equation to ensure convergence with a minimum number of iterations. Importantly, the measurement method is easy to implement on any standard battery cycler with basic math capabilities and does not require any external calculations or analysis to complete. The method can also be applied in situ – during a drive cycle or other operation profile performed on a battery cycler – thereby capturing the impact of recent electrochemical dynamics on power capability. Validation was performed for NMC, NCA, and LiFePO4 chemistries from -20°C to 40°C and for power pulse lengths from 2 s to 30 s, showing accuracy approaching that of the battery cycler. The creation of the measurement procedure was not trivial, requiring hundreds of hours of testing to finalize an approach in which the measured power value was not affected by power pulses applied in prior iterations. This power capability measurement method, the first of its kind, is expected to enable significant improvements to the design, characterization, and management of battery packs.