• On-device AI cuts data traffic and energy use
• CSEM’s Fibonacci chip analyzes audio, motion, and images
• Miromico’s radio supports LoRa® or mioty® connectivity

CSEM and Miromico have developed a compact wireless node that analyzes industrial sensor data at the source. Combining CSEM’s ultra-low-power Fibonacci AI chip with Miromico’s modular radio technology, the platform activates its LoRa® or mioty® connection only when relevant information needs to be transmitted. This reduces data traffic and supports long-term, battery-powered monitoring in remote industrial environments.

AI processing at the sensor

Fibonacci integrates machine-learning accelerators for efficient processing of continuous data streams. It connects directly to sensors such as microphones, accelerometers and small cameras, allowing neural networks to detect anomalies in audio, motion, or image data on-chip. Only relevant results need to be transmitted, conserving energy and bandwidth.

A hardware-optimized compiler maps a range of neural networks onto the chip, enabling one platform to support predictive maintenance, condition monitoring, acoustic monitoring and vision-assisted inspection.

“Fibonacci brings useful AI directly to industrial sensors, where every milliwatt and every transmitted byte matter. By processing data locally, the node can identify relevant events without continuously sending raw information, opening the way to more autonomous monitoring systems,” said Stéphane Emery, Group Leader, Integrated & Wireless Systems at CSEM.

Flexible connectivity for industrial deployment

Miromico’s modular radio handles communication between the device and gateway. The same hardware platform can be configured for LoRa® or mioty®, allowing connectivity to be adapted without redesigning the node.
Designed for sustainable long-term operation

Adaptive power modes align energy use with the workload, while local processing limits radio activity. The result is a compact platform designed for battery-powered monitoring where wired power or continuous connectivity is impractical.

“Industrial customers need sensing solutions that are efficient, adaptable and straightforward to deploy. Combining local AI with a choice of long-range radio technologies gives manufacturers a practical route to monitor equipment in places where power and bandwidth are limited,” said Schekeb Fateh, CEO of Miromico.

Swiss chip design serves local industry

This development highlights how Swiss chip design expertise supports local industry in improving energy efficiency and strengthening data security. CSEM Zurich works closely with local companies to transform world-class semiconductor technologies into deployable solutions, helping to strengthen technological sovereignty in a critical field while enhancing the competitiveness and innovative capacity of Swiss industry.
Further information

Watch the video on our website to discover how CSEM’s Fibonacci AI chip and Miromico’s wireless platform turn sensor data into actionable insights, while reducing data traffic and energy use in demanding industrial environments.