EDATEC’s ED-CLAWBOX is a compact edge AI system based on the Raspberry Pi Compute Module 5. The aluminum desktop system is designed for local OpenClaw deployment and provides Gigabit Ethernet, dual-band Wi-Fi, Bluetooth, dual HDMI outputs, USB 3.0 connectivity, mSATA expansion, integrated audio, and a preinstalled JishuShell management environment.

The Luckfox Lyra PLC is a compact Linux-based programmable logic controller built around the Rockchip RK3506B processor. The DIN-rail system provides dual 100Mbps Ethernet, Wi-Fi 6, Bluetooth 5.2, industrial serial interfaces, CAN, relay outputs, and a 1.9-inch touchscreen.

Seeed Studio’s reComputer Classic J5011 and J5012 are a pair of AI development systems based on NVIDIA Jetson AGX Orin modules. The reComputer Classic J5011 uses the 32GB AGX Orin module, while the J5012 steps up to the 64GB version. Both systems are intended as direct replacements for NVIDIA’s Jetson AGX Orin Developer Kit while retaining compatibility with its peripheral ecosystem.

Norwegian-based Tenko has released an open-hardware development board based on STMicroelectronics’ STM32C5 microcontroller family. The compact board follows the familiar Blue Pill form factor while adding USB-C power, onboard USB-to-UART connectivity, and an alternative firmware environment based on Oberon-2.

Apollo Automation has launched the ESPHome Starter Kit, a modular smart-home development platform built around an ESP32-C6 board and designed to let users create ESPHome devices without soldering, breadboards, or writing code. The kit operates locally and can be used independently or integrated with Home Assistant.

LILYGO’s T-Display SF32 is a compact development platform built around the SiFli SF32LB52X ultra-low-power AIoT microcontroller. The device combines a 2.16-inch 480 × 480 AMOLED touchscreen with dual Arm Cortex-M33 cores, an ePicasso 2.5D graphics engine, LoRa connectivity, Bluetooth 5.3, and onboard battery power.

Shuttle’s XPC slim DH810S is a 1.35-liter barebone PC supporting Intel Core Ultra 200 “Arrow Lake-S” processors with up to 24 cores and a 13 TOPS NPU. The system is intended for digital signage, POS/POI, healthcare, industrial, and edge AI applications.

ILABS is offering the Challenger+ RP2350 NB-IoT, a compact development board combining Raspberry Pi’s RP2350 microcontroller with STMicroelectronics’ ST87M01-1301 cellular module. The board targets low-power IoT applications requiring LTE Cat NB2 connectivity, GNSS positioning, and long battery life.

Radxa has revealed two Qualcomm-based system-on-modules aimed at edge AI, multimedia, robotics, and industrial applications. The new rCore-Q8280 uses the Snapdragon 8cx Gen 3 with 29+ TOPS of AI performance, while the smaller rCore-Q6490 is based on the Dragonwing QCS6490 and provides up to 12 Dense TOPS.

ESP32 Bit Pirate has released version 1.7 of its ESP32-S3-based hardware debugging platform, adding LoRa support for SX1262 transceivers, a Meshtastic analysis shell, a new BPIO2 USB adapter, and support for additional boards. The open-source project continues to expand its browser-based tools for hardware debugging, programming, and experimentation.

Tindie recently listed Comu, a compact development board built around the WCH CH32V203 RISC-V microcontroller. Measuring 13 × 9.4mm, the board fits inside a USB-A port and combines programmable touch controls, LEDs, USB device support, and access to several GPIO signals.

NXP has released the FRDM i.MX 95 Pro, a development board designed for evaluating the company’s i.MX 95 applications processor in edge AI, industrial networking, robotics, vision processing, and advanced HMI applications.

Raspberry Pi introduced a new 3GB version of the Raspberry Pi 4 Model B in April, positioning it between the existing 2GB and 4GB configurations. Four months after the announcement, the model is appearing through additional distributors, including TME, PiShop.us, and Waveshare.

Banana Pi has announced public availability of its BPI-AI2N embedded vision platform, following the hardware introduction covered last year. The platform combines a system-on-module based on the Renesas RZ/V2N Vision AI MPU with a separate carrier board designed for smart cameras, industrial vision, robotics, and other edge AI applications.