Geniatech just announced new SYSTEM-ON-MODULES built around NXP’s latest i.MX95 processor. If you’re into embedded systems and edge AI like me, you know this is a big deal.
The new SOMs, SOM-iMX95-OSM and SOM-iMX95-SMARC, are intended for smart retail, industrial automation, medical devices, and IoT applications. AI is needed, and reliable with long-term support is imperative, and this solves that.
The hw specs are impressive. The i.MX95 chip has 6 Arm Cortex-A55 cores (up to 2GHz! and 2 real-time cores (M7 and M33) completing that mix for big applications with quick reaction to sensor input or control tasks. Plus, NXP embedded its eIQ® Neutron NPU for AI workloads. These modules can do edge vision processing and predictive maintenance (no cloud ping needed). And it operates on 3-8 W.
I like that Geniatech split the lineup into two SOM flavors.
The OSM version (45 x 45 mm) is smaller, tougher, and is soldered directly to the board, making it ideal for smaller and more tightly packed devices with industrial-grade performance.
The SMARC version (82 x 50 mm) provides more room for interfaces and additional expansion options, which is ideal for larger systems, or system upgrades down the line.
The low operating temperatures from -40°C to +85°C, only add to the already tough build of both versions.
When it comes to security, the hardened modules include the NXP EdgeLock secure enclave with encrypted storage support and secure boot support for device safeguarding throughout the device life cycle.
The long-term planning focuses on the NXP 15-year availability program, which Geniatech has been aligned with for some time on the embedded ARM line. This is a significant advantage for prolonged service life equipment like industrial and medical gear that is designed to remain in service for 10 years or more.
These modules seemingly first showed at CES 2026, with sampling shipping in Q1 so I’m definitely looking at getting a dev kit when available. I anticipate it being a solid platform for experimenting with the modern EDGE AI applications.
