Taiwanese IoT firm Advantech has introduced two compact computers-on-module (CoMs) based on the NXP i.MX 95 platform for edge AI applications.

These modules have a multi-domain architecture with up to six Arm Cortex-A55 cores running at up to 2.0GHz, along with a dedicated 2Tops NPU optimised for Linux-based edge workloads. Compared with previous i.MX generations, they deliver up to four times the machine-learning performance, double the data throughput and 15% more power efficiency.

Dutch electronics giant NXP’s (www.nxp.com) i.MX 95 series is the latest addition to its i.MX 9 series of application processors. The i.MX 95 integrates high-performance computing, immersive 3D graphics with Arm Mali, an NXP machine-learning accelerator, and fast data throughput.

This feature set is built to support the growing demands of applications in automotive, industrial automation, networking, connectivity and human-machine interfaces.

As edge intelligence continues to grow, so does the need for powerful processing, real-time machine learning and high-speed connectivity. The i.MX 95 meets these demands as the first in the line to feature its eIQ Neutron neural processing unit (NPU) and an image signal processor (ISP), empowering developers to create smarter, faster and more efficient edge products.

Looking ahead, Advantech’s AOM-5521 (Smarc) and AOM-2521 (OSM) are more than hardware upgrades; they serve as building blocks for edge intelligence. Whether used in industrial automation, medical devices, intelligent vision systems or other IoT applications, these modules can help engineers accelerate development, lower power budgets and expand the possibilities of edge computing.

As edge AI continues to transform industries, Advantech says it is committed to providing agile, scalable platforms that bring tomorrow’s innovation to life today. Learn more about Advantech and its range of RISC products at risc.advantech.com.

Advantech (advantech.com) specialises in IoT intelligent systems and embedded platforms. It focuses on edge computing and edge AI.