Electronics Era: What is the USP of the STM32N6 microcontroller?
Sridhar: The STM32N6 is an exceptional microcontroller that boasts a powerful Arm® Cortex®-M55 CPU operating at 800 MHz, enhanced by Arm Helium vector processing technology for improved DSP performance. The most distinctive feature is the ST Neural-ART Accelerator™, a proprietary neural processing unit specifically intended for power-efficient edge AI applications. These cutting-edge technologies actuate real-time neural network inference for both computer vision and audio applications, thereby making the STM32N6 an excellent choice for those interested in high-performance and AI-driven tasks.
The STM32N6 microcontroller is groundbreaking for two main reasons:
• It features the ST Neural-ART Accelerator, a custom neural processing engine that offers significantly higher machine learning performance.
• It is the most powerful STM32 MCU to date, with a Cortex-M55 running at 800 MHz, high-speed memory interfaces, the largest RAM in an STM32, built-in graphics capabilities, a powerful image signal processor, and an H.264 encoder.
Electronics Era: What trends are driving the next phase of microcontroller innovation?
Sridhar: The industry is increasingly focusing on adding more intelligence, capabilities, and efficiency at the edge. This shift is driven by advancements in microcontrollers, microprocessors, sensors, power devices, and other components, enabling new applications that were not possible a few years ago.
ST is at the forefront of this shift, leveraging more than 10 years of R&D in edge AI and continuously striving to enhance intelligence, capabilities, and efficiency at the edge. Our STM32 ecosystem has been pivotal in making new applications achievable, particularly with innovations like the STM32N6 microcontroller.
Electronics Era: What are key hardware features of the STM32N6? What difference do they bring?
Sridhar: The STM32N6 microcontroller is featured with high hardware specifications. ST Neural-ART Accelerator™ performing real-time neural networks inference, thus capable of supporting even complex AI models at higher level of accuracy. Arm® Cortex®-M55 core with 800 MHz frequency provides the viability of high-performance computing added with DSP, thus improving efficiency for computing. Embedded RAM of about 4.2 Mbytes renders the microcontroller usefully applicable in real-time data processing application and multitasking. Advanced graphics and multimedia include NeoChrom™ Accelerator for 2.5D graphics, H.264 hardware encoder, high-resolution image signal processor (ISP) for image processing. Also coming with strong and robust security features, thus they can easily meet SESIP Level 3 and PSA Level 3 and above benchmarks because they come with the presence of TrustZone® technology for boot and execution management safety for data in addition to integrity of the data.
Key hardware features of the STM32N6 include:
- Cortex-M55 running at 800 MHz
- 4.2 MB of embedded RAM
- High-speed memory interfaces (HexaSPI and OctoSPI)
- NeoChrom GPU for 2.5D graphics acceleration
- Custom image signal processor easily programmable with our free ISP IQTune configuration software
- Security features targeting SESIP Level 3 and PSA Level 3 certifications
- The technology applied for the STM32N6 is 16nm FinFET. This technology allowed us to push the N6 to the target performance, power budget and price point.
Electronics Era: Does ST have AI learning libraries?
Sridhar: Yes, indeed. STMicroelectronics provides AI learning libraries with their ST Edge AI Suite and NanoEdge™ AI Studio. ST already has a growing model zoo. Now we add models to this model zoo that specifically request the performance that the N6 can offer. With that we also engage with external partners to provide turn-key solutions and N6 optimized models. Keep following the N6 for more news on these partners in 2025.
Electronics Era: What future advancements can we expect in AI-driven microcontrollers, and how will they shape the next generation of embedded systems?
Sridhar: ST is providing a full ecosystem of SW tools and partners to design new applications with the STM32N6. We’ve developed a core technology (ST Edge AI Core) to optimize and convert neural networks from popular AI frameworks to take full advantage of the Neural-ART Accelerator. This includes the support of popular AI frameworks like TensorFlow Lite, Keras, and ONNX.
But this is only a small part of what we’re providing. On top of the different tools, we also include the STM32N6 firmware package with computer vision demo applications and an entire suite of software to help build machine learning applications at the edge.
