Today for AI
HOT RADAR
computeHEAT 11.4°

AI CURATED HOT STORY · 10/23/2026

Five Partners Develop 2nm Low-Power AI Chip for Medical Imaging

1 reports archived1 independent sourcesupdated 10/23/2026, 16:59:41
Synthesis & Latest Updates
1 Sources Cross-Validated

LSTC, Hokkaido University, Fujifilm, GE Healthcare, and imec have announced a collaboration to develop a 2nm low-power AI inference chip for medical imaging under the NEDO project. The chip targets real-time reconstruction of photon-counting CT data and intraoperative image recognition, aiming for 5x speed and 10x energy efficiency compared to current GPUs, with initial optimization on IBM's Spyre accelerator.

LATEST/LSTC, Hokkaido University, Fujifilm, GE Healthcare, and imec have announced a collaboration to develop a 2nm low-power AI inference chip for medical imaging under the NEDO project. The chip targets real-time reconstruction of photon-counting CT data and intraoperative image recognition, aiming for 5x speed and 10x energy efficiency compared to current GPUs, with initial optimization on IBM's Spyre accelerator.

HEAT TRENDHourly heat curve

2 fully observed hours
Fewer than 3 fully observed hours; trend not plotted yet.

TIMELINECoverage timeline

Total 1 reports · Latest first
  1. IT之家 · IT业界T2·68 pts
    • Joint development by LSTC, Fujifilm, GE Healthcare, and others under Japan's NEDO project.
    • Target applications include real-time photon-counting CT reconstruction, multimodal image overlay, and intraoperative microscope image recognition.
    • Sets hard metrics of 5x processing speed and 10x energy efficiency over existing GPUs, using a strategy of software simulation before hardware migration.