POLYN Technology Advances Tire Grip Monitoring with Successful NASP Chip Test

POLYN Technology, a fabless semiconductor company specializing in Neuromorphic Analog Signal Processing (NASP), has taped out its first product chip featuring an analog neuromorphic core for a Voice Activity Detection neural network model. The chip aims to qualify the NASP platform and support initial customer products while demonstrating the technology’s performance and robustness.

The NASP technology enables cloudless, on-chip AI for IoT devices, targeting applications in automotive, consumer electronics, industrial IoT, machine-to-human interfaces, and wearables. It uses analog neurons with selective connections to achieve low power consumption and reduced latency.

The chip development involved a neural network compiler and a customized EDA flow using Cadence tools, including Virtuoso for analog design and Innovus for digital implementation, to convert a digital neural network model into analog neuromorphic cores. This process supports scalable neural network layouts with thousands of neurons and resistive weights, compatible with any semiconductor foundry.

POLYN is collaborating with customers on voice processing and sensor data pre-processing for automotive applications. The NASP technology is set for full qualification in Q2 2025, according to CEO Aleksandr Timofeev.