The increasing adoption of Internet of Things (IoT) in critical sectors like healthcare, energy, and industrial systems underscores the need for robust and secure data backup solutions. IoT devices often handle sensitive and mission-critical data, but they are prone to failure due to cyberattacks, hardware failures, and environmental factors. Ensuring end-to-end security and asynchronous recovery in such environments is essential to maintaining operational continuity and data integrity. We address the challenge of securely backing up sensitive data stored in Trusted Execution Environments (TEEs) on IoT devices. We focus on devices that fail without prior knowledge of replacement devices. Such scenarios necessitate secure backup mechanisms that work without requiring interaction between old and new devices. We formulate several architectures that leverage Key Encapsulation Mechanisms (KEM), Proxy Re-Encryption (PRE), and Attribute-Based Encryption (ABE). We address emerging quantum threats by also considering PQC primitives. We evaluate multiple backup schemes regarding security, flexibility, and resilience in a unified testbed. We further validate two approaches through a full proof-of-concept implementation and experimental evaluation using a real-world IoT testbed. Our findings show that secure, scalable, and quantum-resilient IoT backup systems can be built using modular cryptographic approaches. Our comparative evaluation of multiple backup schemes provides crucial insights for designing future-proof IoT Infrastructures, highlighting the trade-offs between different architectural approaches and cryptographic primitives.