InHand Networks launched the Mo 62A and Mo 68A AI single-board computers on July 28, 2026, targeting edge vision AI applications in industrial equipment, access control systems, robotics, smart cameras, and visual inspection terminals. The boards ship with a preconfigured development environment that reduces the setup and integration work typically required before teams can start building functional prototypes.
"Edge AI product development often involves more than running a model," said a spokesperson for InHand Networks. "Developers also need to connect cameras, process video, deploy models, handle inference results, and integrate those results into real device workflows. The Mo 62A and Mo 68A are designed to help teams build this foundation more efficiently and focus more effort on model performance, application logic, and product experience."
What's included in the development environment
The boards come with Debian, TI EdgeAI SDK / TIDL, OpenCV, GStreamer, and other commonly used components necessary for camera input, video processing, on-device inference, and result output. This preconfigured stack lets teams skip weeks of foundational configuration and move directly to model testing and integration.
Hardware and AI performance
- Preconfigured development environment: The Mo 62A and Mo 68A support the full vision AI workflow-camera input, image processing, inference, and result handling-without requiring separate toolchain setup.
- Edge vision workflow support: Both boards accept MIPI CSI camera input, integrate a 7th-generation ISP, and include H.264 / H.265 video codecs. This covers the pipeline from raw video capture to processed output.
- Flexible AI performance: The Mo 62A provides 2 TOPS for lightweight single-camera recognition tasks. The Mo 68A delivers 8 TOPS for multi-model processing, multi-stream video analysis, industrial visual inspection, intelligent transportation, and robot vision.
- Field deployment readiness: Hardware watchdog, RTC timekeeping, secure boot, TrustZone, OP-TEE, and hardware encryption give the boards a foundation for long-term unattended operation in industrial, security, and embedded systems.
The Mo 68A is designed for heavier workloads that involve multiple models or higher video throughput. InHand Networks notes that actual video stream capacity and model concurrency depend on model complexity, resolution, frame rate, frame sampling strategy, and application logic.
Why this matters for product development teams
The boards condense what is often a fragmented setup process into a single platform. Instead of spending weeks getting cameras, inference engines, and toolchains to work together, product development teams can start with a working pipeline and focus on the parts that differentiate their product-model accuracy, application logic, and the end-user experience. The inclusion of secure boot and remote deployment features also means the same hardware can move from early prototyping into field trials without a redesign.
Your membership also unlocks:
