At the STM32 Summit last December, ST innovation leaders announced the groundbreaking AI-accelerated MCU, the STM32N6, presenting inspiring customer use cases, and sharing the stage with our partner Edge Impulse. They also explored our strategic collaborations with Qualcomm Technologies, Inc, and Amazon Sidewalk in the field of wireless applications.
What do you need to know about the STM32N6? Let’s summarize the main takeaways.
1. An advanced AI microcontroller with an NPU
The STM32N6 microcontroller is built on a powerful Arm® Cortex®-M55 processor clocked at 800 MHz, the highest STM32 core frequency yet. It features an ST proprietary neural processing unit (NPU), enabling extended neural network computing capabilities within the power budget of a microcontroller. The Neural-ART Accelerator is one of the most powerful accelerators in this range of devices seen in the industry until now.
2. High processing power
The STM32N6 microcontroller offers 600 GOPS of processing power, 3 TOPS per watt power efficiency, and provides 4.2 Mbytes of embedded RAM for real-time data processing and multitasking. It also supports H.264 encoder and JPEG encoder/decoder and features a 2.5D graphics accelerator, enhancing multimedia capabilities.
3. Image signal processor dimensioned for 5 Mpixel cameras at 30 FPS
The STM32N6 microcontroller is among the early microcontrollers to combine an image signal processor (ISP) with a neural processing unit (NPU). The ISP optimizes images directly on the microcontroller, enabling cost savings thanks to the use of simpler image sensors. Developers can use the free STM32-ISP-IQTune tuning tool to easily enhance CMOS image quality.
4. An extensive ecosystem to fast-track development
- Discovery kit: the STM32N6570-DK includes a 5" LCD module with capacitive touch panel, a 5-megapixel camera, an IMU, a Time-of-Flight sensor, user buttons, an SD card slot, an audio jack, USB connection, Gbyte Ethernet, and a Raspberry compatible camera connector.
- Essential AI software tools: the STM32N6 comes with a dedicated collection of tools and resources to simplify the development and deployment of AI models. For instance, the STM32 model zoo offers a collection of optimized neural networks, categorized by application types like object detection, pose estimation, and audio scene recognition, simplifying the training and deployment of AI models. Visit this webpage to access software application examples, and go the ST Edge AI Suite to find all the software for running AI on edge devices.
Other STM32N6 dedicated events will be announced soon. Stay tuned to secure your spot!
First published on Jan 29, 2025
