Many emerging high-volume sensing applications are, by necessity, wireless. Thus, each sensing and communication node must contain its own power source or be able to harvest its own power. Example applications include smart buildings and homes, wearable health and wellness sensors (i.e. mobile health), structural health monitoring, smart transportation environments, biomedical implants and much more. Today, almost all wireless sensors use batteries as their sole power source. In some cases this is not limiting. In other applications, where recharging batteries is a nuisance or expensive, energy harvesting is clearly beneficial. In other applications, regular battery replacement is impossible and thus energy harvesting is critical. This presentation will focus specifically on wearables and harvesting energy from human body motion to power them. There are two fundamental challenges with harvesting energy from human: very low frequency excitation and multi-directional excitation. Because of the low frequency excitation, linear (as opposed to rotational) energy harvesters are usually displacement limited. Rotational architectures exist (i.e. self-powered watches) but their power production is exceptionally low, much too low for current and envisioned wearables. We have developed models to predict the maximum possible power that could be generated from regular human motions given realistic size constraints. In this presentation, I will discuss the implications of these models for product design and demonstrate that much more power is available than is produced by existing products. We have developed a series of prototypes to enhance the power output using both electromagnetic and piezoelectric transducers. These prototypes perform significantly better than the current state of the art, but are still well below the theoretical limit. Thus, there is still much room for improvement and innovation.