TDK develops analog AI chip that mimics brain function for edge computing

TDK, a company best known for audio cassettes, has developed a prototype analog reservoir AI chip that mimics brain function for real-time learning applications. The chip, created in collaboration with Hokkaido University, uses analog circuitry to process time-varying data at high speed and ultra-low power, making it suitable for robotics and human-machine interfaces that require instant feedback.

How it works: The chip mimics the human cerebellum and uses natural physical dynamics of analog signals for efficient processing.

• Unlike traditional deep learning models that rely on cloud processing and extensive datasets, this technology learns directly at the edge using analog circuitry for reservoir computing.
• The silicon uses wave propagation and other analog signal dynamics to interpret input and produce output with minimal power consumption.
• Its real-time learning capability allows rapid adaptation to changing data streams, making it ideal for wearable devices, autonomous systems, and IoT hardware.

The demonstration: TDK will showcase the chip at CEATEC 2025 with an interactive rock-paper-scissors game that predicts player moves.

• The demo device attaches to users’ hands and uses acceleration sensors to track finger movement.
• The AI chip processes this movement data in real-time to predict the winning gesture before players complete their move.
• “In rock-paper-scissors, there are individual differences in finger movement, and in order to accurately determine what to do next, it is necessary to learn those individual differences in real time,” TDK explained.

The bigger picture: This development represents TDK’s evolution from magnetic materials into advanced neuromorphic computing.

• The new design builds on earlier TDK research into neuromorphic devices that attempted to mimic the cerebrum using spintronics.
• Instead of handling heavy computational jobs, this analog reservoir AI focuses on quick, low-power processing of time-series data for sensing and control applications.
• TDK hopes the demonstration will “foster a broader understanding of reservoir computing” and accelerate commercialization of reservoir computing devices for edge AI applications.

What’s next: TDK plans to extend its collaboration with Hokkaido University and apply the results to its Sensor Systems Business and TDK SensEI brand.