This document discusses energy theft detection issues in smart grid advanced metering infrastructure (AMI). It presents an attack tree threat model showing energy theft behaviors in AMI. Currently, meeting demand is not possible and energy theft in distribution cannot be found. The proposed system uses current sensors to detect power theft, sends alerts via GSM, and manages data using Zigbee. This prevents electricity theft and the associated problems of power crisis and load shedding. Hardware includes microcontrollers, sensors, GSM/Zigbee modems, and software is Kiel Compiler and embedded C. A block diagram shows the system components and connections. Future uses include industries, hospitals, and power suppliers to address the estimated $25 billion lost annually to energy theft