After 50 years of Moore’s law, it’s now possible to pack nearly 1 million transistors in the space of a paramecium, enabling tiny systems for sensing, communication, and computation. This radical miniaturization of electronics has brought with it the incredible opportunity to reimagine life in the microworld: put together the right way, these parts could be used to build robotic creatures too small to be seen by the naked eye. This talk is about building microscopic robots. I’ll show how a new class of nanometer thick, electrochemical actuators can be used to build mechanical components that are controllable with electronic signals, opening the door to integrating silicon electronics with small-scale moving parts. As proof of concept, I’ll show a simple robot that walks using on-board silicon photovoltaics for power. Each step in fabricating these robots is carried out massively in parallel, allowing millions of electronically integrated robots to be built on a single 4-inch wafer. Next, I’ll present ongoing work in our lab to build a programmable robot just over a hair’s width in size, complete with memory, a controller, sensors, and on-board power. Looking forward, I’ll argue that tiny robots offer a new platform to explore physics in the microworld, ranging from swarms and collective behavior to low-Reynolds number locomotion. Finally, I’ll discuss early work on applications, including a project to regrow damaged nerves by literally pulling them where they need to go with tiny machines.