AAEON Unveils Software Guardian for Edge AI Algorithms

Comprised of server-side management tools, secure tunnel communication, and device-level security mechanisms, AAEON’s new security package aims to protect AI models on the edge.

(Taipei, Taiwan – August 13, 2025) AAEON, an industry-leading provider of edge AI solutions, has unveiled a new software security platform tailored for its line of embedded AI systems powered by NVIDIA® Jetson™ system-on-modules. The security platform, which is built on a three-tiered architecture with components designed to protect data both at the edge and the cloud, is now available as part of the board support package for SKUs of AAEON’s BOXER-8621AI, BOXER-8641AI-Plus, and the BOXER-8651AI-Plus Edge AI systems.

The most notable component of the software security offering is a Trusted Execution Environment (TEE) named ‘MAZU’, which is designed to protect AI models and application data by segmenting different files, processes, and algorithms within protected execution zones. By leveraging MAZU, users will be able to isolate machine learning application algorithms within a protected “Secure World”, while allowing standard applications to run in a “Normal World.” As a consequence, access to sensitive assets will be gated through certified APIs with encrypted communications, secure OS environments, and certificate validation.

Other device-side mechanisms included in the new framework include Anti-Restore Protection, A/B Redundancy partitioning, and Disk Lock. These components prevent hackers from reverting system software to previous versions, allow the system to revert to a stable OS image in the event of device failure, and encrypt physical storage located on edge devices to protect sensitive data at rest, respectively.

The package contains a number of server-side management tools to assist with the management of edge devices at the server level, including a web-based UI through which multiple edge systems can be monitored and managed from a single server, while this layer also provides more robust over-the-air (OTA) update protocols. Meanwhile, data transmitted between the server and edge devices will be protected via a secure tunnel, allowing users to safely implement key tasks such as AI model updates.

Given the company’s recent introduction of its AI Inference Server product line, which include the MAXER-2100 and MAXER-5100, the integration of an ecosystem-focused security framework serves as a statement of intent, hinting that AAEON’s focus lies not only in the provision of NVIDIA-accelerated hardware platforms, but also auxiliary offerings built to protect device fleets across entire edge applications.

Detailed information about AAEON’s software security framework can be found in the company’s recently published white paper, while pricing and ordering information for product SKUs equipped with the platform are available via the AAEON website.