Aisin Unveils World's First 16nm AI SoC with Integrated STT-MRAM, Boosting Efficiency 50x

1 reports archived1 independent sourcesupdated 11/3/2026, 09:30:09

Synthesis & Latest Updates

1 Sources Cross-Validated

Aisin and Tohoku University have developed the world's first 16nm AI SoC integrating Spin-Transfer Torque MRAM (STT-MRAM). Utilizing a Near-Memory Computing architecture to store model weights directly on-chip, the solution reduces boot time significantly and achieves a 50x improvement in energy efficiency for object detection.

LATEST/Aisin and Tohoku University have developed the world's first 16nm AI SoC integrating Spin-Transfer Torque MRAM (STT-MRAM). Utilizing a Near-Memory Computing architecture to store model weights directly on-chip, the solution reduces boot time significantly and achieves a 50x improvement in energy efficiency for object detection.

Hourly heat curve

3 fully observed hours

Now

11.4

Peak

11.910:00

24h change

–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first

IT之家 · IT业界T2Nov 3, 09:30·68 pts
Aisin Unveils World's First 16nm AI SoC with Integrated STT-MRAM, Boosting Efficiency 50x
- World's first 16nm AI SoC integrating 32MB STT-MRAM, enabling local storage of model weights and system data.
- Adopts Near-Memory Computing (NMC) design, reducing boot-to-OS time to 65.7ms with ~38x speed improvement.
- Achieves 50.7mJ/frame energy consumption for Tiny YOLOv3 via hybrid multiplication and weight sharing, a 98% reduction compared to traditional solutions.
View Related Reports & Analysis Original Source