Nordic Semiconductor is acquiring Neuton.AI’s core technology and IP. Neuton.AI specializes in automating Tiny ML processes, making it easy for non-experts to build AI models. Nordic plans to integrate Neuton.AI’s platform with its own nRF54 Series chips. These are chips that use very little power, with popularity among battery-operated devices. The combination should allow for smarter devices without the need for more hardware.
Neuton.AI’s models are tiny, usually under 5 kilobytes, and fast, contrasting with traditional AI models that are large and slow. In addition, the models are efficient in terms of energy and memory, and use very little space on a chip, making them all the more ideal for constrained devices. Manual tuning or data science expertise is not needed in their system, meaning developers can train and deploy models without having to understand complex machine learning theory. The tech offers compatibility across both old and new products, supporting a wide array of microcontrollers (8-, 16-, 32-bit).
Consumer electronics, healthcare, and industrial applications are the markets in focus. The deal includes a team of 13 engineers and data scientists, who developed and maintain the TinyML platform. Current users of Neuton.AI will remain disruption-free during the integration. This move positions Nordic to tap into the TinyML market, where chip sales could reach $5.9 billion by 2030, according to market forecasts.
