Highlights

  • EMASS prepared ECS-DoT demonstrations covering asset tracking, voice-enabled vision and predictive maintenance for electronica 2026.
  • The Bosch Sensortec collaboration combines five sensor types with local AI processing for event-driven asset monitoring.
  • A new reference design combines face detection and keyword recognition within 512 KB of on-chip memory.
  • Nanoveu shares rose 11.11% to AUD 0.050 on 10 November 2026.

Battery-powered devices that can interpret more than one type of signal are central to the latest semiconductor work at Nanoveu (ASX:NVU). Through its EMASS business, the company has developed demonstrations that process visual, audio and environmental information close to the sensor. The program connects a July collaboration with Bosch Sensortec, an October exhibition announcement and a technical demonstration detailed in early November.

Several Sensor Inputs, One Chip

EMASS announced in October that it would present three ECS-DoT applications at electronica 2026 in Munich, scheduled for 10–13 November. These included intelligent asset tracking, vision-activated voice control and predictive maintenance. The presentation was intended to show how an ultra-low-power system-on-chip can interpret different kinds of information without relying on continuous cloud processing.

For asset tracking, its jointly developed design with Bosch Sensortec combines motion, environmental, magnetic, pressure and gas sensors. Instead of maintaining constant GPS and wireless activity, the architecture is designed to activate higher-power functions when locally processed signals indicate a meaningful event. Its possible uses include cold-chain shipments, Warehouse logistics, industrial equipment and high-value goods.

Bosch Collaboration Creates Evaluation Platform

The July collaboration produced a reference design available for evaluation by customers building battery-powered trackers. It is intended to provide more context than a location reading alone: movement, handling and surrounding conditions can inform decisions about an item in transit. The design operates within a coin-cell battery power budget.

The development represents a reference architecture rather than a mass-production contract. Developers can adapt the design to different operating environments, with commercial deployment remaining a separate stage of development.

Vision And Audio Work Together

A technical update published on 6 November described a single ECS-DoT running a face-detection model and a voice-command model. The camera first checks whether someone is present. Only then does the microphone open for recognised commands, including “up”, “down”, “left”, “right” and “go”. Images and audio remain on the device in this reference design.

The ECS-DoT operates with a 50 MHz core and 512 KB of on-chip memory, with the complete demonstration occupying 447,392 bytes. The face detector took approximately 13.8 milliseconds per inference; keyword inference took 174 microseconds. In a test using 114 video frames, the face detector identified 72 of 74 faces and produced no false alarms across 40 frames without a face. These measurements describe the demonstrated configuration.

Next Steps Focus On Wider Multimodal Processing

EMASS intends to progress from two alternating AI models toward a single multimodal model combining sound and visual inputs in one inference. Its immediate exhibition plans also include showcasing equipment monitoring using vibration, motion and environmental signals. Both initiatives illustrate the company’s stated objective of supporting always-on intelligence within constrained power budgets.

Nanoveu shares gained 11.11% to AUD 0.050 on 10 November 2026. The company had a Market Capitalisation of AUD 54.37 mn.