The AstraEdge showcase demonstrates a live end-to-end workflow that links GenAI reasoning with physical robotic execution and real-time digital twin validation, powered by NVIDIA MGX servers, Jetson Thor, and Isaac Sim.

Demos will feature the ECA-6051 server within the NVIDIA MGX architecture, equipped with L40S GPUs, driving Lexa—the agentic AI engine that processes complex, interactive voice commands via large language models in real time—and the EAI-I351 Robotic AI Computer using Jetson Thor and Vision-Language Models to control a robotic arm.

AstraEdge is positioned as enabling autonomous systems to understand, plan, and act within a unified, high-performance edge computing environment, signaling a broader shift toward industrial edge computing.

Lanner Electronics is a long-standing provider of edge AI platforms and network appliances with a 30-year track record and an ecosystem of partnerships.

The architecture targets latency-sensitive environments such as warehouses and factories, enabling on-site perception, reasoning, and real-time insights while preserving data sovereignty.

Overall takeaway: Aetina positions itself as a key infrastructure provider for industrial and enterprise autonomy, bridging high-precision sensing with edge AI cognition to accelerate real-world deployment of physical AI technologies.

The AIB-AT78/68 supports up to 2070 FP4 TFLOPS and includes high-speed connectivity (QSFP28, EtherCAT, dual 10 GbE) to address sensor fusion and motion control needs, leveraging NVIDIA Isaac workflows and Vision AI.

Availability note: AIB-AT78/68 series is expected to be commercially available in late March 2026, with live demonstrations at Booth 139 during GTC 2026.

In collaboration with Stereolabs, the Humanoid Surround Vision Solution will use two ZED X stereo cameras and two ZED X One cameras to achieve sub-millisecond 3D perception, operating from -25°C to 80°C and delivering 130W of power efficiency.

The AstraEdge portfolio aims to bring data center–class AI performance to the edge to enable low-latency, real-time autonomous decision-making where data is generated.

Lanner emphasizes that AstraEdge delivers data center–class AI performance at the extreme edge to support rapid, localized decision-making in edge environments.

Lanner describes AstraEdge as bringing data center–class AI capabilities to the extreme edge to enable low-latency, real-time decisions at the data source.