The objective of the proposed research is to develop a thermoelectric energy harvester CMOS power-supply system. Thermoelectric generators (TEGs) can power and sustain biomedical implants and IoT microsensors in medical and industrial applications where light and motion are inaccessible. Millimeter-scale TEGs, however, are resistive, intermittent, low-voltage, and as a result, also low-power. An energy-harvesting system for this type of imperfect source must not only track the maximum power point (MPP) but also include and replenish a small battery, regulate and supply a load, wake from no-charge conditions, and coordinate and control several intertwined feedback loops. Unfortunately, state-of-the-art TEG systems are largely incomplete. Some can wake and track the MPP, but not include or replenish a battery or supply a load, or the other way around. And when tracking the MPP is included, it is often suboptimal. This research proposes to develop a CMOS power-supply system that can track the MPP, replenish and draw assistance from a small battery, supply and regulate a load, and wake from no-charge conditions. Two fundamental goals for this research are to develop an accurate MPP model that can be used to optimize the system and build a harvester that outputs the highest maximum power possible. One way to improve the performance of the system is by improving the efficacy of the switched-capacitor circuit used to wake the harvester. Another way is to design circuit components so they duty-cycle and require little quiescent current to operate. With all this in place, IoT microsensors and biomedical implants in dark and remote areas can sense, process, and transmit information that can save money, energy, and lives.