Geniatech’s new SOM lineup based on the NXP i.MX95 application processor signals a strong step forward in industrial edge computing. The company’s latest releases — the SOM‑iMX95‑OSM and SOM‑iMX95‑SMARC — combine high compute density, integrated AI acceleration, and robust industrial‑grade reliability in two practical form factors.
Both modules center on the i.MX95’s 6 × Arm Cortex‑A55 cluster (up to 2.0 GHz) supplemented by dual real‑time cores (M7 + M33). This heterogeneous architecture allows deterministic control and multitasking for industrial and automation environments where timing precision matters. Geniatech further leverages NXP’s eIQ® Neutron NPU for on‑device inference, supporting vision analytics, edge classification, and predictive tasks without relying on cloud backhaul.
From a systems‑integration perspective, expandability is well covered: USB 3.0, PCIe Gen 3, CAN‑FD, dual TSN‑capable GbE, and MIPI‑CSI/DSI display interfaces. Multimedia features include a 4K video pipeline and Arm Mali GPU, which position these modules for use in HMI terminals or AI inspection systems.
Security and safety compliance remain consistent with industrial expectations — NXP’s EdgeLock secure enclave supports chain‑of‑trust boot, hardware key storage, and encrypted OTA updates, aligning with ISO 26262 ASIL B / IEC 61508 SIL 2 requirements.
Form‑factor choice depends on deployment strategy:
SOM‑iMX95‑OSM (45 × 45 mm) targets solder‑down, size‑constrained designs where mechanical stability and vibration resistance are critical.
SOM‑iMX95‑SMARC (82 × 50 mm) offers greater I/O availability and upgrade flexibility for modular systems and gateway‑class hardware.
Each unit supports operation from –40 °C to +85 °C and comes with Geniatech’s full BSP, Yocto support, and long‑term lifecycle assurance under NXP’s 15‑year program — a strong commitment for customers in automation and medical segments.
The modules debuted at CES 2026, with evaluation kits scheduled for availability in Q1. For integrators exploring next‑gen industrial edge computing platforms, this looks like a well‑balanced entry point that blends compute power, longevity, and maintainability.
