[Exhibition] VIA to Unveil Next-Generation Edge AI Platforms at Japan IT Week Spring 2026

VIA Technologies Japan Inc. will exhibit at "Japan IT Week Spring 2026," held at Tokyo Big Sight starting Wednesday, April 8, 2026.

At this exhibition, under our new identity "Visvia," which signifies VIA's evolution into the next generation, we will showcase numerous practical edge AI solutions powered by the two major platforms: MediaTek Genio and NVIDIA Jetson.

We invite you to experience the latest solutions at the VIA booth, which solve common challenges in AI field implementation such as "cloud communication and API costs," "the burden of selecting and developing hardware suitable for field environments," and "the hassle of integrating with peripheral devices." These solutions smoothly support customers in transitioning their AI projects from PoC (Proof of Concept) to mass production.

1. [Automotive & Fleet Management] "D800" & "J600" for Rapidly Building Custom Fleet Management Systems

We will exhibit an in-vehicle AI solution that has cleared vibration resistance standards (ISO 16750-3) for commercial vehicles and construction machinery that require high environmental durability.

The next-generation high-end dashcam "Mobile360 D800," making its debut in Japan, not only comes standard with predictive danger AI but also provides an AWS integration SDK. It functions as a platform for SIers and service providers to quickly develop their own "original connected dashcam service." Additionally, we will exhibit the "Mobile360 J600," an 8-inch rugged terminal with built-in AI that provides a 360° surround view.

2. [Manufacturing & Visual Inspection] "AI Architecture Comparison" and a Line-Integrated "Dedicated Edge AI Inspection Machine"

To assist in selecting the optimal hardware according to manufacturing line requirements, we will conduct two demonstrations.

The first is a "Visual Inspection Architecture Comparison Demo" using NVIDIA Jetson Orin NX (AMOS-9100) and MediaTek Genio 700 (VAB-5000). We will directly run an AI model developed in Python to demonstrate how differences in hardware architecture (GPU vs. NPU) affect the AI approach. This goes beyond mere spec comparisons to offer new system design options and ideas for the inspection challenges our customers face. The second is a live demonstration of a "Dedicated Edge AI Inspection Machine" based on the AMOS-9100. It integrates interfaces from camera input to PLC and patrol light control into a single fanless, rugged chassis, significantly reducing the system integration effort required for setting up inspection equipment and lines.

3. [Retail & Security] "Edge Generative AI" with No API Fees and a High-Precision "Facial Recognition Terminal"

For retail KIOSKs and smart customer service, we will demonstrate "Edge Generative AI (Edge LLM)" that operates completely offline without using the cloud. This eliminates network latency and the need for pay-per-use API costs.

For access control, we will exhibit the "ACS-5000" access control terminal, which features an RGB + IR (infrared) dual-lens camera to block spoofing attempts with photos or videos at the hardware level. As it has been verified to work with major facial recognition software like FaceMe, it can be quickly integrated into existing systems.

Exhibition Overview