Boards Guide
Maker's Guide to 2025 Boards: Exploring the Future of Microcontrollers & SBCs
Get ready for an exciting year with "The Guide to Boards 2025"! This edition shines a spotlight on specialization, featuring boards meticulously crafted for the cutting-edge fields of AI, IoT, and LED projects. Notable innovations include the AI-driven BeagleBoard BeagleY-AI and Seeed Studio’s Vision AI modules, perfect for makers interested in machine learning and automation.
The Arduino Nano family shines as a versatile option, covering wearables, IoT, and classroom projects. From the Nano 33 IoT with built-in Wi-Fi and Bluetooth to the classic Nano’s compact form factor, these boards excel in wireless communication, motion tracking, and seamless integration with Arduino ecosystems, making them a go-to for both beginners and seasoned makers.
For smart environments, the M5Stack Dial offers a unique rotary input and RFID reader, ideal for home automation projects. LED enthusiasts can look to Pimoroni Plasma boards, which now include options powered by the RP2350 microcontroller, delivering easy control of RGB strips with robust I/O options. Meanwhile, the DFRobot LattePanda Mu brings PC-level performance to SBCs with its Intel N100 processor, making it a powerhouse for both Windows and Linux applications.
DigiKey's Innovation Hub is an immersive web experience that lets you explore featured products, tools, boards, calculators, and more. Designed for students, makers, and engineers, it serves as your one-stop shop for a carefully curated selection of resources and content that DigiKey proudly offers. The Innovation Hub encompasses a Test Bench, Boards Explorer, Featured Products, Calculator, and Knowledge Center. Click below to visit the Innovation Hub's Board Explorer and uncover all the details on boards highlighted in the 2025 Make: Guide to Boards!
BEAGLEBOARD BEAGLEY-AI
One of the first exciting new releases of 2024 was the open-source AI-focused BeagleY-AI from BeagleBoard. A departure from their typical “cape” headers, the BeagleY adopts a Pi-like form factor, with quad 1.4GHz 64-bit Arm Cortex-A53 cores, a powerful GPU supporting three simultaneous displays, and a built-in 4 TOPS AI co-processor.
View ProductUSEFUL SENSORS PERSON SENSOR
Another AI-in-a-box solution is the Person Sensor from Useful Sensors (not to be confused with their product literally named “AI in a Box”). This thumbnailsized board uses its built-in camera and smarts to detect and identify faces, without requiring users to obtain a degree in machine learning. In addition to making it simple to use, the abstracted black-box functionality also helps ensure privacy.
View ProductPIMORONI PLASMA 2350
This issue’s cover focuses on RP2350-based boards,and Pimoroni are a leading adopter of RPi’s new silicon. Their Plasma board is a USB-C powered controller for addressable RGB LED strips, with screw terminals for easy attachment, a Qwiic connector for convenient expansion, and 3A of juice to power fun glowies like their optional 10-meter LED star strip.
View ProductOPENMV CAM RT1062
OpenMV are the “Arduino of Machine Vision” and their latest-gen smart camera board features the same pioneering MicroPython-based implementation with an updated 600Mhz i.MX RT1060 Arm Cortex-M7 and built-in Wi-Fi. Compatibility with over a dozen existing shields and a modular camera system makes theirs the ideal board for a range of AI projects.
View ProductSEEED STUDIO GROVE VISION AI MODULE V2
On the smaller side, Seeed’s Grove Vision AI v2 Kit bundles an AI-optimized Arm Cortex-M55/EthosU55-based board with an optional OV5647 Camera Module and Xiao ESP32-C3 to create an approachable machine vision solution leveraging no-code SenseCraft pre-trained models. An extensive wiki and examples help makers take things to the next level with their own custom AI solutions.
View ProductCROWPANEL ESP32 1.28" ROUND DISPLAY
Elecrow have been cranking out a series of ESP32 + display combos under the CrowPanel moniker, and it was hard to pick just one, but in the end this stunning round capacitive touch model won out. Ideal for wearables, home automation, and IoT projects, it also supports the fantastic LVGL graphics library.
View ProductM5STACK DIAL
Another spin on the ESP32 + 1.28" round touchscreen formula, this time from the incredibly innovative M5Stack. What makes this one stand out is the rotary dial that surrounds the screen, providing a unique and charming user input method, which combined with an RFID reader and integrated M5StampS3 makes it great for home automation projects where temperature or light levels require convenient adjustment, as well as user identification and authorization.
View ProductDFROBOT LATTEPANDA MU
The LattePanda Mu is DFRobot’s take on the compute modules offered by the likes of Raspberry Pi. Unlike the Arm-powered Pi CM4, however, it is based on the quad-core 3.4GHz Intel N100, with claimed performance of more than double the latest Pi 5, as well as 8GB of LPDDR5 RAM and 64GB of integrated eMMC 5.1 storage. An open-source carrier implementation turns it into a tiny PC motherboard, complete with PCIe slots, perfect for your next Windows or Linux-based project.
View ProductKHADAS VIM1S
We featured the VIM4 in an earlier Guide, and now we’re excited to see Khadas grace these pages again with the VIM1S. Its combination of compact size, yet full-size ports and built-in storage (saving you a bundle on specialty cables and accessories), plus great graphical performance and Oowow OS delivery service make it stand out in the single board computer (SBC) space.
View ProductRASPBERRY PI 5
An accessible, low-cost single-board computer ideal for edge AI applications. Its powerful CPU can run full machine learning frameworks like PyTorch and TensorFlow, and supports GPU acceleration for inference (not training). Object detection (YOLOv8n) for images with 640×640 resolution runs about 5 FPS, and small LLMs (Llama8:3B) can generate responses at about 2 tokens per second. USB and PCIe slots allow for add-on AI acceleration hardware, which can dramatically increase speeds.
View ProductNVIDIA JETSON ORIN NANO
Very powerful SBC with integrated Nvidia GPU for fast inference. Object detection (YOLOv8n) for 640×640 images runs about 30 FPS, and small LLMs (Llama8:3B) can respond at about 4 tokens per second. Training is also feasible, but slow. Much more expensive ($500+) than the Pi, and the Nvidia-specific frameworks present a longer learning curve. Also has USB and PCIe slots for acceleration hardware.
View ProductSEEED STUDIO XIAO ESP32S3 SENSE
Compact, powerful stack of boards featuring a camera, mic, Wi-Fi/BLE radio, and AI acceleration for small tasks such as audio, vibration, and image classification. It can’t handle full object detection (e.g. YOLOv8n), but can use Edge Impulse’s FOMO model for constrained object detection at about 8 FPS, 96×96 resolution.
View ProductRASPBERRY PI PICO 2
The new RP2350 microcontroller is a big upgrade from the RP2040. Its dual-core Cortex-M33 offers special hardware optimizations like SIMD, DMA, and an FPU, for processing time-series data and inference tasks such as audio, vibration, and image classification. While it lacks the ESP32-S3’s connectivity, the Pico 2 is cheaper and should perform similarly well.
View ProductLILYPAD ARDUINO USB
LilyPad Arduino is the original flavor of sewable circuit boards, developed by Dr. Leah Buechley and SparkFun Electronics. Since 2008, the original LilyPad design has inspired many other circular, hand-sewable boards. LilyPad Arduino USB is the latest version, featuring an ATmega32U4 microcontroller chip with built-in USB support, an on/off switch, and a JST connector for LiPo batteries.
View ProductADAFRUIT FLORA
Flora V3 is Adafruit’s dedicated wearable electronics board, and the smallest full-featured one around. Flora has the same ATmega32U4 chip with USB support, plus beginner-proof power management that accepts a versatile range of 3.5V to 16V DC. A full-color NeoPixel LED is perfect for providing visual feedback when you’re debugging your wearable on the go.
View ProductADAFRUIT CIRCUIT PLAYGROUND EXPRESS
CPX is the PCB equivalent of a Mary Poppins bag: Just when you think you’ve explored all the features on this board, you discover there are more! While not intended for wearables, it’s currently by far my favorite all-in-one board for wearables and e-textiles. Packed with on-board sensors and actuators, its powerful ATSAMD21 Arm Cortex-M0 processor can be programmed in (at least) three different ways (see page 7). Connections aren’t limited to conductive thread; other methods demonstrated in the Adafruit Learning System include a bolt-on kit that doesn’t require any soldering!
View ProductARDUINO NANO 33 IOT
New Arduino Nano boards are released almost every year, and the newer ones have become favorites for classroom kits involving wireless communication. The Arduino Nano 33 IoT has been a go-to in recent years, featuring the ATSAMD21 processor, both Wi-Fi and BLE, and an on-board accelerometer/IMU for motion tracking. For wearables, the headerless version works well in small spaces and it can be powered by a power bank via USB.
View ProductADAFRUIT FEATHER FAMILY
With their extensive variety of both main boards and accessory “wing” boards, working with Feathers feels like building a circuit with a PCB card deck of superpowers. Processor choices include ATmega32u4, M0, M4, ESP32, RP2040, and the new RP2350; radios include Wi-Fi, Bluetooth, GSM cellular, LoRa, and more. Short stacking headers, doubler/tripler prototyping boards, and Stemma QT connectors are all options for packing a lot of functionality into a small, robust package, perfect for fitting in a case or pocket of a wearable.
View ProductADAFRUIT FEATHER RP2040 SCORPIO
The Adafruit Feather RP2040 Scorpio board provides eight pairs of level-shifted signal pins with ground pins to easily control eight LED strips at once. Combined with Adafruit’s code library (C++ or CircuitPython), the RP2040’s PIO processors can quickly push LED signal data to eight NeoPixel strips simultaneously while freeing the main processor for other tasks. As part of Adafruit’s Feather ecosystem, the Scorpio is compatible with dozens of add-on boards, and while you must code your own LED patterns, helpful examples on Adafruit’s website will get your project running quickly.
View ProductPIMORONI PLASMA
Pimoroni Plasma boards come in RP2040, RP2040W, and RP2350 varieties and feature screw terminals with two level-shifted data pins for quickly reusable connections to 3-pin and 4-pin LED strips. The USB-C connector provides up to 3A of current at 5V to the LEDs. Two user buttons and lots of I/O pins make it easy to connect peripherals, and the speedy RP2040 and RP2350 processors have plenty of power to run intensive computations (C/C++ or MicroPython) while still animating the LEDs.
RP2040 RP2040W RP2350