Home » Markets » AI

Embedded: Grinn claims AstraSOM-261x as “world’s smallest Edge AI” SoM

Grinn, the Polish embedded Systems-on-Module (SoM) specialist, is highlighting its Grinn AstraSOM-261x at Embedded World.

Embedded: Grinn claims AstraSOM-261x as "world’s smallest Edge AI" SoM

Based on the Synaptics Astra AI-native processor, Grinn describes the product as the “world’s smallest Edge AI” SoM. It is aimed at industrial environments, and has an operating temperature range of -40°C to +85°C.

“The Grinn AstraSOM-261x represents a step forward in our idea to simplify complex electronics,” said Robert Otręba, CEO of Grinn. “By packing the power of the Synaptics Astra 2610 into a tiny 25x25mm module, we are enabling our customers to build smarter, smaller, and more efficient devices. This is the ultimate tool for those looking to a compact solution for making decisions at the Edge.”


Read all our Embedded World content »


AstraSOM-261x

Grinn claims AstraSOM-261x as "world’s smallest Edge AI" SoM - Block diagram

Highlighted features of the AstraSOM-261x include:

  • Single/dual-core Arm Cortex-A55, Cortex-M52, and a dedicated 1 TOPS NPU for AI acceleration.
  • Mali-G31 GPU supporting Full HD @ 60 fps with OpenGL ES 3.2 and Vulkan 1.3 support.
  • Operates on a single 4.5 – 5.5 V DC input and can power external components directly from the module.
  • Built-in low-power modes to extend the life of battery-operated devices.
  • Includes MIPI-CSI/DSI, Dual Ethernet (TSN/1588), USB 2.0, CAN-FD, and high-speed serial interfaces.
  • Enterprise Security: PSA Level 3 Root of Trust, Secure Boot, and on-chip crypto engines.

Synaptics

As the name suggests, the Grinn AstraSOM-261x marks another collaboration between Grinn and Synaptics. It combines Grinn’s compact hardware design with Synaptics’ Astra AI processor family, the two companies are making advanced Edge AI technology more accessible to engineers and manufacturers worldwide.

The Astra SL2610 processors are powered by the new Synaptics Torq™ Edge AI platform, which combines future-ready NPU architectures with open-source compilers to redefine IoT AI development. Torq introduces the first production deployment of Google Research’s Coral NPU, featuring dynamic operator support to future-proof Edge AI designs. By leveraging an open-source IREE/MLIR compiler and runtime, it provides a truly developer-first ecosystem.

You can read more information on the Grinn website. And you can find the Grinn company stand at Nuremberg at 4.260.

Note that a dedicated industrial SBC (single board computer) based on the Grinn AstraSOM-261x is already available, for prototyping.

Images: Grin

Alun Williams

Alun Williams

Web Editor of Electronics Weekly, he is the author of the Gadget Master and Electro-ramblings blogs and also covers space technology news. He has been working in tech journalism for worryingly close to thirty years. In a previous existence, he was a software programmer.

Leave a Reply

Your email address will not be published. Required fields are marked *

*