What you’ll learn:
- Which five new modules target edge AI, industrial IoT, and embedded designs.
- How each platform differentiates itself through NPUs, connectivity, software support, or size.
- Where these modules fit in applications such as smart vision, HMI, smart retail, robotics, and medical systems.
Here’s a look at five more modules that took the spotlight at Embedded World 2026. With an eye toward edge AI and IIoT solutions, these platforms bring a range of differentiating features to the table.
- Variscite VAR-SMARC-MX8M-PLUS
- congatec conga-SMX95 / aReady.COM conga-SMX95
- SECO SOM-Trizeps-X-Genio360/360P
- Grinn GenioSOM-360
- Octavo OSD62x-PM
Variscite VAR-SMARC-MX8M-PLUS
The VAR-SMARC-MX8M-PLUS is a high-performance board for edge AI use in industrial, IoT, and medical settings. It’s one of the first capable of complex machine learning, such as object recognition, thanks to its TOPS 2.3 NPU.
Variscite believes this module serves as the standard for smart vision and industrial automation via pin-to-pin scalability. In addition, the company made the VAR-SMARC-MX8M-PLUS future-proof as it has a TMP 2.0 security chip, dual Gigabit Ethernet, and Wi-Fi 6. Developers can seamlessly advance their projects from prototypes to an industrial rollout without interruptions.
congatec conga-SMX95 / aReady.COM conga-SMX95
The Conga-SMX95/aReady.COM conga-SMX95 from congatec is a high-performance module that focuses on application-ready hardware. While it comes with the NXP i.MX 95 CPU, 2 TOPS eIQ Neutron NPU and LPDDR5 memory, the advantage lies in the aREADY.COM ecosystem integration. A secure OS, software stacks, and a hypervisor enable quicker development cycles.
According to congatec, this module can reduce the time-to-market, especially for medical and industrial settings relying on security. Engineers won’t have to focus on low-level driver debugging for weeks. Instead, they can work on unique AI and edge computing.
SECO SOM-Trizeps-X-Genio360/360P
The SECO SOM-Trizeps-X-Genio360/360P is a powerful module for high-performance edge AI and multimedia purposes. Manufacturers are able to upgrade it down the line according to performance needs, as the module uses an industry-standard SODIMM-200 connector. Developers can leverage the 5.1-TOPS Genio 360 and 8+-TOPS Genio 360P to elevate their system’s AI capabilities.
SECO said this module is well-suited for industrial HMI and smart retail, as 4K60 video output and DisplayPort 1.4 fit that purpose for industrial reliability.
Grinn GenioSOM-360
GenioSOM-360 is a small module featuring MediaTek’s GenioSOM-360 processor. With a 30- × 30-mm form factor, Grinn’s module is designed for physical AI applications running on Cortex-A76 cores and a 7.4-TOPS NPU. The vibration-resistant board serves as a middle ground for low-power IoT and high-end computing.
The company also developed it to last 10 years and run small generative edge-AI models. The GenioSOM-360 module can be deployed in commercial and industrial settings that require small hardware.
Octavo OSD62x-PM
Octavo Systems’ OSD62x-PM is an extremely small module, approximately the same size as a postage stamp. Engineers can use it to design and develop extremely compact, custom industrial equipment without complex high-speed DDR routing. Octavo refers to this board as a game-changer for industrial IoT and robotics, where space is limited. It’s also highly reliable with features like TSN Ethernet and dual 1080p display support, suiting it for rugged applications.
>>Check out Parts 1 and 2 of this series, as well as more of our overall 2026 Embedded World coverage
About the Author
Cabe Atwell
Technology Editor, Electronic Design
Cabe is a Technology Editor for Electronic Design.
Engineer, Machinist, Maker, Writer. A graduate Electrical Engineer actively plying his expertise in the industry and at his company, Gunhead. When not designing/building, he creates a steady torrent of projects and content in the media world. Many of his projects and articles are online at element14 & SolidSmack, industry-focused work at EETimes & EDN, and offbeat articles at Make Magazine. Currently, you can find him hosting webinars and contributing to Electronic Design and Machine Design.
Cabe is an electrical engineer, design consultant and author with 25 years’ experience. His most recent book is “Essential 555 IC: Design, Configure, and Create Clever Circuits”
Cabe writes the Engineering on Friday blog on Electronic Design.
Voice Your Opinion!
To join the conversation, and become an exclusive member of Electronic Design, create an account today!
Leaders relevant to this article:
Facts Only
Variscite VAR-SMARC-MX8M-PLUS: A high-performance board for edge AI use in industrial, IoT, and medical settings. Features TOPS 2.3 NPU, TMP 2.0 security chip, dual Gigabit Ethernet, Wi-Fi 6, and pin-to-pin scalability.
Conga-SMX95/aReady.COM conga-SMARC: A high-performance module for application-ready hardware. Comes with NXP i.MX 95 CPU, 2 TOPS eIQ Neutron NPU, LPDDR5 memory, aREADY.COM ecosystem integration, secure OS, software stacks, and hypervisor.
SECO SOM-Trizeps-X-Genio360/360P: A powerful module for high-performance edge AI and multimedia purposes. Uses an industry-standard SODIMM-200 connector, 5.1-TOPS Genio 360, 8+-TOPS Genio 360P, 4K60 video output, DisplayPort 1.4, and is suitable for industrial HMI and smart retail applications.
Grinn GenioSOM-360: A small module featuring MediaTek’s GenioSOM-360 processor. Has a 30 x 30 mm form factor, Cortex-A76 cores, 7.4 TOPS NPU, and designed for physical AI applications running on low-power IoT and high-end computing.
Octavo OSD62x-PM: An extremely small module, approximately the same size as a postage stamp. Designed to simplify industrial IoT and robotics with features like TSN Ethernet, dual 1080p display support.
Executive Summary
Embedded World 2026 showcased five new modules for edge AI, industrial IoT, and embedded designs. These platforms are designed to improve machine learning capabilities, connectivity, software support, and reduce time-to-market in various applications like smart vision, HMI, smart retail, robotics, and medical systems.
The Variscite VAR-SMARC-MX8M-PLUS is a high-performance board for edge AI use cases in industrial, IoT, and medical settings, with TOPS 2.3 NPU for complex machine learning tasks such as object recognition. The Conga-SMX95/aReady.COM conga-SMX95 from congatec offers a secure OS, software stacks, hypervisor, and quicker development cycles for medical and industrial settings relying on security.
The SECO SOM-Trizeps-X-Genio360/360P is a powerful module for high-performance edge AI and multimedia purposes, suitable for industrial HMI and smart retail applications. The Grinn GenioSOM-360 is a small module designed for physical AI applications running on Cortex-A76 cores and a 7.4-TOPS NPU, while the Octavo OSD62x-PM is an extremely small module (postage stamp size) that simplifies industrial IoT and robotics design with TSN Ethernet and dual 1080p display support.
Full Take
Steelman: The article presents an overview of five new modules that took the spotlight at Embedded World 2026, each designed to target edge AI and industrial IoT solutions with various features such as NPUs, connectivity, software support, or size. These platforms are expected to play a crucial role in applications like smart vision, HMI, smart retail, robotics, and medical systems by improving performance, reducing time-to-market, and enhancing security.
Patterns Detected: None
Root Cause: The development of these modules reflects the increasing demand for edge AI and industrial IoT solutions in various industries, requiring advanced hardware capabilities to support complex machine learning tasks, high-performance multimedia applications, and secure data processing.
Implications: These new modules aim to address the need for efficient, compact, and powerful solutions for edge AI and industrial IoT applications, potentially accelerating innovation and improving business efficiency in industries such as manufacturing, healthcare, retail, and robotics.
Bridge Questions: How will these new modules impact the development of edge AI and industrial IoT solutions? Which industries are most likely to benefit from their adoption, and what challenges may arise during implementation? What additional features or improvements can we expect to see in future iterations of these modules?
Sentinel — Human
The analyzed article appears to be written by a human journalist, with a natural writing style and distinct authorial voice. However, stylometric signals alone can sometimes be misleading in determining machine authorship, so the likelihood of synthesis remains relatively low.