IN Brief:
- Aetina will showcase physical AI, vision AI, and agentic AI edge platforms at COMPUTEX 2026.
- Demonstrations include Jetson Thor robotics, RTX PRO 6000 Blackwell server acceleration, GMSL2 stereo vision, and on-premise AI systems.
- The portfolio reflects growing demand for local reasoning, multi-sensor perception, and industrial AI deployment without cloud dependency.
Aetina will showcase its latest edge AI platforms at COMPUTEX 2026, with demonstrations spanning physical AI, vision AI, and agentic AI for robotics, factory automation, enterprise workflows, and industrial perception.
The exhibit will run from 2 to 5 June at Taipei Nangang Exhibition Center and will cover Aetina’s SuperEdge, MegaEdge, DeviceEdge, and CoreEdge portfolios. The company will show systems developed with partners including NVIDIA, Franka Robotics, Axis, Innodisk, PrefactorTech, Stereolabs, Digital Media Professionals, Cambrian Robotics, Universal Robots, Qualcomm, MoBagel, and Data Systems.
The centrepiece is an AI-powered robotic arm system running at the edge. It combines Aetina’s DeviceEdge AIB-AT78/68, based on NVIDIA Jetson Thor, with the company’s SuperEdge AEX-2UA1 GPU server accelerated by an NVIDIA RTX PRO 6000 Blackwell Server Edition GPU. The demonstration uses an on-premise private LLM and NVIDIA ACE to enable real-time voice interaction through a digital avatar, while a Stereolabs ZED camera provides object recognition and 3D perception.
The physical response is handled by a Franka Research 3 robotic arm and Axis dexterous hand, giving the system a full edge workflow across voice, vision, reasoning, and motion control. Aetina is also showing a humanoid surround-vision system using Jetson Thor, Stereolabs ZED X stereo cameras, ZED X One cameras, and GMSL2 connectivity for low-latency depth sensing and peripheral awareness.
For electronics and automotive production, Aetina will demonstrate an automated wiring and cable assembly system using its MegaEdge AIP-KQ67 inference platform. Cambrian’s AI vision perception guides a dual-arm Universal Robots system to recognise, guide, and insert flexible cables and connectors, targeting a labour-intensive process that is difficult to automate with conventional fixed tooling.
The vision AI section includes a lightweight VLM application running on the palm-sized DeviceEdge AIE-CN33/43-A1 Mini system, powered by NVIDIA Jetson Orin NX 16GB. Aetina will also show the Arm-based DeviceEdge AIE-KR1B-B1, whose M.2 modular architecture supports accelerator cards from suppliers including Axelera AI, DEEPX, EdgeCortix, MemryX, and Realtek.
The COMPUTEX showcase follows a steady rise in Jetson Thor industrial platforms. Advantech has already launched Jetson Thor edge AI systems, while Lanner has targeted rugged autonomous systems with Jetson Thor compute. Aetina’s emphasis adds another layer: hardware is being shown as part of deployable robotics and workflow systems, not simply as accelerator capacity.
That shift is important for industrial AI adoption. Factory systems need local data control, predictable latency, rugged I/O, thermal stability, model deployment tools, and maintainable hardware roadmaps. As physical AI moves from demonstration to production, the edge platform has to behave like industrial infrastructure — not a lab workstation with a robot attached.
