Many recent computational models of the brain address “functional” features of neuronal activity – that is, information-processing patterns of units in the system, treated as abstracted function of stimulus input or a time variable. However, the real brain is a physical device embedded in space, exhibiting reliable spatial organization, strongly constrained by biophysical requirements, and subject to substantial size, weight, and power limitations. Recent work in NeuroAI has begun to address these key facts, leading to an array of exciting theoretical modeling approaches to the brain as a biophysical system; exposing a set of new and unsolved empirical questions; and enabling a spectrum of potentially high-impact real-world neural applications. This symposium will focus on each of these components, including cutting-edge presentations on theory, experiment and application; and across a spectrum of brain areas and systems.