STMicroelectronics announces a new generation of microcontrollers integrated with NPU accelerators to help edge artificial intelligence development

     STMicroelectronics introduces a new family of microcontrollers with integrated machine learning (ML) accelerators to bring embedded artificial intelligence (AI) to life, enabling cost - and power-focused consumer electronics and industrial products to run algorithms such as computer vision, audio processing, sound analysis, and more. It provides high performance functions that previously small embedded systems could not achieve.

     The STM32N6 series of microcontrollers (MCUS) is ST's product with high processing performance to date, and it is also the first microcontroller embedded with ST's self-developed Neural network processing unit (NPU) neural-ART Accelerator. Machine learning processing performance is 600 times higher than the existing high-end STM32 MCU products. The STM32N6 will be available to selected customers from October 2023 and is now ready for mass market distribution.

     Remi El-Ouazzane, President of ST Microcontrollers, Digital ics and RF Products (MDRF), said: "The micro edge device market is poised for a sea change, with more and more devices adopting AI models to improve the performance of the master controller or take over the work of the master controller. These models are currently used for tasks such as segmentation, classification, and identification, and will enter new applications that have not yet been developed in the future. The STM32N6 is the first STM32 product to use ST's self-developed Neural-ART Accelerator NPU, and the version of our self-developed AI software ecosystem package will support STM32N6, which is the beginning of a long road of development for STM32 with AI hardware accelerator. It will enable application product innovation that is not possible with any other embedded processing solution."

     Tom Hackenberg, memory and computing analyst at Yole Group, said: "Previously, the general view was that AI was just a big data center, a very power-hungry application. That view is no longer tenable. The STM32N6 is a great example of the new trend of integrating the energy-efficient processing power of microcontrollers with the analytical power of AI, which is urgently needed in today's iot edge applications, to handle computer vision and many sensor-driven applications, while delivering low power and cost to the user's products."