Arduino UNO Q Explanation: Hybrid Board With Linux Computer + Microcontroller Explained for Those Curious About It

Arduino UNO Q Explanation: Hybrid Board With Linux Computer + Microcontroller Explained for Those Curious About It Meta description: Arduino's new UNO Q board is a departure from the usual Arduino board, combining a full Linux computer and a real-time microcontroller on a single board, blurring the lines between an "Arduino project" and a "Raspberry Pi project". Introduction Arduino has released one of their most exciting boards in years, the new UNO Q which aims to bring a full Linux computer and a real-time based microcontroller onto a single board, blending the distinction between "Arduino project" and "Raspberry Pi project". What Makes the UNO Q Different Arduino boards have always been microcontrollers, which is excellent for real-time based hardware projects, but don't offer much in terms of an operating system and processing power for anything more involved than basic logic. The UNO Q board changes this by combining two different p...
Post a Comment

Edge AI on Microcontrollers - What This Means for Makers in 2026

Edge AI on Microcontrollers - What This Means for Makers in 2026

Meta description: AI is coming to microcontrollers, not just the cloud. See what this means for Arduino, ESP32, and Raspberry Pi makers in 2026.

Introduction

If you've been following embedded development for more than three years, you have an intuitive grasp of the long-standing paradigm of AI: you send data to the cloud and receive (usually much slower) data back from the cloud. What most makers don't realize is that this paradigm is set to change in 2026. The new crop of microcontrollers in various development boards are being engineered to actually perform AI functions on-device, and this has huge potential for both hobbyists and industrial applications alike.

In short, this will let you do interesting edge AI/ML/TinyML things without having to purchase and maintain a server or even a Raspberry Pi - at least not right away.

This article will discuss what this means for makers and hobbyists and how it impacts the boards that we already own or are considering purchasing.

Let's dive in!

Background and What's Really Changing

Traditionally, microcontrollers were never designed to perform "AI" functions like object detection or speech recognition. That was always left to much more powerful, energy-intensive, and expensive devices like a Raspberry Pi. Even if you had a modest AI/ML-capable microcontroller, you needed some serious amount of RAM and storage to perform meaningful AI functions. This changed in 2025 - with the advent of manufacturers producing microcontrollers with onboard capabilities to actually perform AI functions. Some manufacturers (like TI, STMicro, NXP, and Renesas) even have entire product lines dedicated to this concept now, and they're all set to release even more exciting products in 2026.

In short, we have both the silicon and the ecosystem to support "edge AI"/TinyML on microcontrollers now.

This has several implications for makers and hobbyists.

Specific Implications for Makers

ESP32 already has a head start here. Espressif's newer ESP32-S3 family of System-on-Chips have vector instructions built-in that can be leveraged to perform certain ML/AI operations much more efficiently than a general-purpose CPU. This is already a great way to experiment with edge AI with the ESP32 without having to get a Raspberry Pi or even a dedicated AI/ML-capable microcontroller.

Raspberry Pi's position in the ecosystem will change. While the Raspberry Pi has traditionally been the go-to choice for any AI/ML edge device, its position will increasingly be supplanted by AI/ML-enabled microcontrollers. Yes, it will still serve a purpose for heavy lifting, but for simpler edge-AI applications, a microcontroller could be a much more power-efficient and cost-effective choice.

Novel applications will begin to appear. For example, there are several applications where performing even rudimentary "on-device AI" could improve functionality or enable an entirely new class of applications for existing devices. These could range from detecting wake-words to enable/disable devices, to in-sensor analytics to detect specific patterns of interest.

TinyML means doing ML/TinyML on microcontrollers. What does this really mean?

TinyML stands for running trained machine learning models on devices with very limited resources, such as microcontrollers. A TinyML model is generally smaller (often on the order of 1MB or even less), more optimized (to use less memory and be more power-efficient), and more specific (to accomplish a narrower set of tasks) than your average ML model. It's not a complete AI/machine learning stack but a single, optimized, trained model that does one specific thing very well.

Edge AI is when you do AI functions, such as ML or TinyML, on an edge device like a microcontroller rather than in the cloud.

How can I get started with Edge AI?

There are several ways to get started with Edge AI, depending on what you already have:

  • If you have an ESP32, you can try out TensorFlow Lite for Microcontrollers
  • If you have a Raspberry Pi or other board with an optional add-on AI accelerator, you can try out Edge Impulse (very good for getting started)
  • Or you can try out TensorFlow or PyTorch models on a Raspberry Pi with an optional add-on AI accelerator

I'm looking forward to trying out TensorFlow Lite for Microcontrollers on my existing ESP32, now that I have one of the appropriate variant (ESP32-S3).

Will I need special hardware to start playing around with Edge AI or TinyML

Not necessarily. Some of the more modern ESP32 devices allow you to run very simple TinyML models directly. Having a microcontroller that has dedicated silicon for this will undoubtedly make it faster and more power-efficient, but it's not strictly necessary to get started.

Is Edge AI just about running ChatGPT-style models on a microcontroller?

Not really. Edge AI and TinyML are generally about deploying very specific and narrow ML models. ChatGPT or similar models are "large language models" (LLM) and require a lot of memory and processing power to run. These are generally only found on cloud servers or in devices that have specialized silicon to run LLMs.

Should I purchase additional hardware to start experimenting with TinyML or Edge AI?

If you already have an ESP32 (ideally the S3 variety) or a Raspberry Pi, you can absolutely experiment with Edge AI and TinyML right now without purchasing anything else. That said, if you're interested in getting deeper into this space, you might want to purchase some of the more modern AI/ML-oriented microcontrollers. The performance and battery life improvements can be substantial, especially on resource-constrained devices like wearables or robotics.

Do I have to stick with what I already have or can I get microcontrollers that are more suited to AI?

You absolutely can get microcontrollers that are better suited for AI functions. Some microcontroller manufacturers have specific product lines dedicated to this, and these generally offer significantly better performance and power efficiency for AI workloads. This will doubtless become more prevalent in the coming years as this becomes a mainstream thing.

Summary

Edge AI will affect the devices that we use in our daily lives and hobbies. As microcontrollers that are capable of (or optimized for) Edge AI become mainstream, we can expect to see various novel applications of this technology in both our work and hobbies. I expect there to be more tutorials about TinyML and Edge AI on this blog as this ecosystem evolves.

Comments

Post a Comment

Popular posts from this blog

ESP32 DHT22 Temperature & Humidity Sensor Tutorial

ESP32 DHT22 Temperature & Humidity Sensor Tutorial (With Code & Wiring Diagram) Meta description: Learn how to connect a DHT22 sensor to an ESP32 and read temperature and humidity data with this complete wiring diagram, code, and troubleshooting guide. Introduction The DHT22 is one of the most popular sensors for measuring temperature and humidity in DIY electronics and IoT projects. Paired with an ESP32 — a microcontroller with built-in Wi-Fi and Bluetooth — it's a great starting point for weather stations, home automation systems, or any project that needs environmental data. This guide walks through wiring the DHT22 to an ESP32, writing the code to read sensor data, and displaying temperature and humidity readings on the Serial Monitor. What This Project Does You'll build a simple circuit where the ESP32 reads temperature (in Celsius and F...
Read more

ESP32 Web Server | Display DHT22 Temperature & Humidity in Your Browser

ESP32 Web Server: Display DHT22 Temperature & Humidity in Your Browser Meta description: Turn your ESP32 into a local web server that displays live DHT22 temperature and humidity readings in any browser on your network — full code included. Introduction If you've already wired up a DHT22 sensor to your ESP32 (see our ESP32 DHT22 tutorial), the next logical step is getting rid of the Serial Monitor and viewing your readings on an actual web page — from your phone, laptop, or any device on your network. In this tutorial, you'll turn your ESP32 into a self-hosted web server that displays live temperature and humidity data from the DHT22 sensor directly in your browser. What This Project Does The ESP32 connects to your Wi-Fi network, hosts a simple web page, and refreshes temperature and humidity readings from the DHT22 sensor automatically. No inte...
Read more

Arduino HC-SR04 Ultrasonic Sensor Tutorial. Measure Distance with Code & Wiring Diagram

Arduino HC-SR04 Ultrasonic Sensor Tutorial: Measure Distance with Code & Wiring Diagram Meta description: Learn how to wire and program an HC-SR04 ultrasonic sensor with Arduino to measure distance in centimeters and inches — full code and troubleshooting included. Introduction The HC-SR04 ultrasonic sensor is one of the most widely used components in beginner Arduino projects. It measures distance by sending out a sound wave and timing how long it takes to bounce back — the same basic principle used by bats and sonar. In this tutorial, you'll learn how to wire the HC-SR04 to an Arduino, program it to measure distance, and display accurate readings in both centimeters and inches using the Serial Monitor. What This Project Does You'll build a simple distance-measuring circuit where the Arduino triggers the HC-SR04 to send an ultrasonic pulse, measures the time ...
Read more