Efficient adaptation of large language models (LLMs) on edge devices is crucial for applications that need ongoing, privacy-preserving adaptation and inference. This is particularly crucial for mission-critical applications, such as drones, where hardware resources are limited, and the system’s overall security must operate within a zero-trust environment. In order to address these challenges, we propose a novel low-overhead hardware architecture that enables secure LLM operations on the edge. Furthermore, the proposed scheme incorporates a secure communication framework to maintain the integrity of edge operations against an external zero-trust network environment. If successful, this project will lay the foundation for a low-overhead edge environment to execute LLMs while maintaining the privacy and integrity of the system.