The low gravimetric and volumetric energy densities of energy storage technologies are principal limitations for realizing the full potential of modern robotics technologies[1]. In many cases, the size and weight of energy storage technologies required to power robotic systems are too large or massive for a robot to carry untethered. Therefore, most robotic systems are tethered, or constrained to limited operational times and long-recharging times over which they remain unused. Energy storage technologies are particularly limiting for micrometer to centimeter scale robots, vehicles, and sensors, as microbatteries dominate the size and mass of the corresponding devices [2,3]. One approach to overcome the energy limitations of batteries is to use battery chemistries with increased energy density. Metal-air batteries are particularly appealing due to their high energy densities. Iron-air, zinc-air, and aluminum-air batteries have 1229, 1535, and 8076 Wh/kg theoretical energy densities [4], which is 5 – 32 X larger than commercial lithium ion batteries. Practical metal-air batteries, however, only achieve a fraction of the theoretical energy density (350 – 500 Wh/kg for Zn-air, for example) and suffer from poor and inefficient rechargeability. Although promising, realizing rechargeable metal-air batteries with near theoretical energy densities is likely several years or decades away. New approaches for powering robots that take advantage of specific environments or use-cases are required to realize untethered robotic systems in the near future. In a conventional battery, active materials are stored in a rigid package so that the maximum volumetric energy density is the energy available in the lowest capacity electrode divided by the total battery volume or mass. A robot, therefore, is required to carry both the active material mass as well as the packaging and electrolyte required for the battery to work. In metal-air batteries, the pertinent electrode is the solid metal anode which oxidizes to produce electrons and, typically, a metal hydroxide or metal oxide. An electrolyte connects the metal anode to a thin porous cathode that reduces oxygen in the air. Metal-air batteries have high energy densities because oxygen for the cathode reaction can be extracted from the air and therefore does not contribute to the battery mass or volume. Electrons from the anode are transferred to the cathode through the robot electronics, powering the robot. In this work, we present a metal-air scavenger (MAS), which is a device that extracts electrical energy from external metal surfaces to power centimeter scale or smaller robots, vehicles, and electronics. The MAS undergoes the same chemical reactions as a metal-air battery, but is only composed of a polymer electrolyte and cathode current collector so that both the anode and cathode active materials are external to the device. The powered devices, therefore, only need to carry the mass of the electrolyte and cathode current collector. We show that, when stationary, a MAS extracts near the theoretical capacity of aluminum and zinc surfaces, as well as extracts up to 1,450 mAh/g from iron surfaces. The maximum power densities of a MAS on aluminum, zinc, and iron surfaces are 130, 80, and 25 mW/cm2. A principal advantage of the MAS is that it can continue to power a device as it moves across a metal surface, so that the total mass and volume of metal oxidized are many multiples of the mass and volume of the device. We demonstrate the utility of a MAS by powering a 5 x 4 x 2 cm electric vehicle on aluminum and zinc surfaces with a 2 x 2 x 0.3 cm MAS. The MAS powered the vehicle over many square feet of metal surface and at up to 8 mm/s velocity. The MAS concept presented here takes advantage of the prolific variety of external metal energy sources available in modern urban environments. We believe the MAS will find utility in many applications, lead to further unconventional methods for powering robotic systems, and open new materials development opportunities. Yang, Guang-Zhong, et al. "The grand challenges of Science Robotics." Science Robotics3.14 (2018): eaar7650.St. Pierre, Ryan, and Sarah Bergbreiter. "Toward Autonomy in Sub-Gram Terrestrial Robots." Annual Review of Control, Robotics, and Autonomous Systems(2019).Oudenhoven, Jos FM, Loïc Baggetto, and Peter HL Notten. "All‐solid‐state lithium‐ion microbatteries: a review of various three‐dimensional concepts." Advanced Energy Materials1.1 (2011): 10-33.Li, Yanguang, and Jun Lu. "Metal–air batteries: will they be the future electrochemical energy storage device of choice?." ACS Energy Letters2.6 (2017): 1370-1377. Figure 1