Lanner Electronics will unveil its next-generation AstraEdge™ portfolio at NVIDIA GTC 2026, positioning edge infrastructure as a critical enabler for generative AI-driven robotics and telecom networks. The company plans to demonstrate how edge systems can connect large-scale AI reasoning with real-world robotic execution and digital twin validation, a workflow increasingly viewed as essential for the next phase of industrial automation.
The showcase reflects a broader shift underway across data centers and telecom networks. Enterprises now deploy generative AI models not only for analytics but also for real-time operational control. Lanner’s AstraEdge™ lineup addresses this shift by bringing data-center-class AI compute directly to the 5G extreme edge, where latency and responsiveness shape the success of automation systems.
The company will introduce a portfolio of edge platforms designed to support generative AI robotics, AI video analytics and AI-RAN deployments. These systems integrate NVIDIA MGX architecture, NVIDIA Jetson Thor modules and NVIDIA Isaac Sim environments, creating a unified infrastructure layer for reasoning, execution and validation across distributed edge locations.
“At NVIDIA GTC 2026, Lanner is demonstrating how Generative AI can move beyond digital intelligence to real-world action, combining scalable Edge AI servers and rugged Edge AI platforms from NVIDIA to enable autonomous robotics, video analytics, and AI RAN at the edge.”
AstraEdge portfolio targets the extreme AI edge
The AstraEdge portfolio arrives as industries move away from rule-based automation toward autonomous reasoning systems. Manufacturers, logistics operators and telecom carriers increasingly demand infrastructure capable of interpreting complex instructions, interacting with physical environments and validating outcomes in real time.
Consequently, edge infrastructure must process advanced AI models without relying on centralized cloud environments. Lanner’s approach centers on integrating high-performance GPUs, robotics-focused AI modules and simulation frameworks into compact edge systems. This architecture allows AI workloads to run close to sensors, machines and network endpoints, significantly reducing inference latency.
At GTC 2026, Lanner will present a live demonstration that connects generative AI reasoning with robotic execution through a closed-loop workflow spanning three specialized systems.
The reasoning layer operates on the ECA-6051 Edge AI Server built on NVIDIA MGX architecture and equipped with NVIDIA L40S GPUs. The system runs “Lexa,” Lanner’s agentic AI engine, which processes interactive voice commands using large language models in real time. Lexa converts conversational instructions into structured commands that downstream robotics systems can execute.
Meanwhile, the execution layer relies on the EAI-I351 Robotic AI Computer powered by NVIDIA Jetson Thor. The platform applies Vision-Language Models and Vision-Language-Action models to interpret visual context and translate AI reasoning into physical actions. In the demonstration, a robotic arm responds to commands generated by the Lexa system, illustrating how generative AI can coordinate physical operations.
Finally, the validation layer operates through the EAI-I730 Edge AI Workstation, which supports both the NVIDIA RTX PRO 4500 Blackwell Server Edition GPU and RTX PRO™ 6000 Blackwell Server Edition. Running NVIDIA Isaac Sim, the workstation creates a real-time digital twin environment where robotic behavior can be simulated, validated and synchronized with physical hardware. This closed-loop model testing allows engineers to analyze outcomes instantly and refine AI behavior before large-scale deployment.
AI-native RAN infrastructure also enters focus
Beyond robotics, Lanner will also introduce the ECA-6710, a scalable edge AI server designed to support NVIDIA ARC-Compact architecture. The system targets telecom operators transitioning toward AI-native radio access networks as the industry prepares for 5G-Advanced and early 6G architectures. Telecom infrastructure increasingly relies on AI workloads for network optimization, traffic prediction and autonomous resource allocation. Therefore, edge servers capable of hosting AI inference directly within RAN environments have become essential for operators seeking both performance gains and operational efficiency.
By aligning edge computing with generative AI and telecom infrastructure, Lanner’s AstraEdge strategy signals a broader convergence shaping the digital infrastructure landscape. Robotics, industrial automation and telecommunications now share the same AI-driven compute foundations.
At GTC 2026, Lanner’s demonstration aims to show how these domains increasingly intersect at the edge, where AI reasoning, machine control and real-time simulation converge to power the next generation of autonomous systems.
