Specifications
Platform
NVIDIA Jetson Thor
Board Type
Medical AI board / industrial system board
Power
130W total — no additional GPU modules required
AI Workloads
Real-time surgical instrument anomaly detection, organ segmentation, AR overlay
Software
Advantech edge AI software tools + medical AI frameworks
Imaging Pipeline
Low-latency imaging pipeline for intraoperative video
Memory
Jetson Thor unified LPDDR5X memory (up to 128GB class)
Deployment
Medical imaging devices, surgical systems, intraoperative guidance
Form Factor
Embedded board for integration into OEM medical equipment
Certification Path
Medical-grade system integration with Advantech USM-500 / IGX platforms
Showcased
NVIDIA GTC 2026 — real-time medical imaging AI
Differentiator
GPU-class imaging AI on 130W, without a discrete GPU card
Overview
The Advantech AIMB-294 is a medical AI board built on NVIDIA Jetson Thor that brings GPU-class imaging AI to surgical and diagnostic equipment without a discrete graphics card. Advantech demonstrated it at GTC 2026 performing real-time surgical instrument anomaly detection, organ segmentation and augmented-reality overlay in a 130W envelope. Because Jetson Thor integrates CPU, GPU and memory in a single system-on-module, the AIMB-294 removes the PCIe graphics card, its airflow requirements and its power delivery from the medical device bill of materials — a significant simplification for cart-based and intraoperative equipment where space, heat and acoustic noise are constrained.
Key Benefits
130W replaces a GPU card: the entire imaging AI workload runs on the Jetson Thor SoM, so OEMs can drop the discrete GPU, its cooling and its power supply from the design. Real-time intraoperative performance: transformer and segmentation models run at clinical video rates on live imaging feeds, enabling AR overlay and instrument tracking during a procedure rather than as a post-processing step. Lower integration risk: a standard embedded board form factor plus Advantech's edge AI software tools and medical AI frameworks shortens the path from prototype to a certified device. One silicon roadmap: the same Jetson Thor platform and JetPack stack scales across imaging products, so a manufacturer can carry software investment from a flagship surgical system down to a compact ultrasound unit.Applications
Intraoperative imaging and surgical guidance systems, surgical instrument tracking and anomaly detection, organ segmentation for planning and navigation, AR overlay in operating theatres, ultrasound and endoscopy AI, and portable or cart-based diagnostic imaging devices.
Request a Quote — ADVANTECH AIMB-294 — MEDICAL AI BOARD ON NVIDIA JETSON THOR, 130W
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →