en.Wedoany.com Reported - ASRock Rack Inc. unveiled its latest AI-native infrastructure product portfolio at COMPUTEX 2026, designed for the agentic AI era. This includes the new 2UXGM-VERA2 system powered by the NVIDIA Vera CPU, showcasing next-generation AI platforms for AI factories, inference cloud deployments, and edge AI applications.

The AI industry is undergoing a fundamental transformation, with scaling laws moving beyond model size into the era of agentic AI. This shift makes CPU execution a critical path. ASRock Rack's 2UXGM-VERA2 system features the Vera processor, which combines custom-designed NVIDIA Olympus cores, high-bandwidth LPDDR5X memory, and the NVIDIA scalable coherence fabric to deliver a unified computing architecture. It can execute complex agentic and reinforcement learning workloads up to 50% faster than infrastructure using traditional CPUs.

Weishi Sa, President of ASRock Rack, stated that the AI industry is entering a new frontier where systems shift from generating responses to executing autonomous actions. The NVIDIA Vera is a CPU for the AI era, transforming model inference into autonomous action. ASRock Rack plans to launch a comprehensive new line of AI servers based on NVIDIA Vera, ranging from standalone CPU servers to scalable AI servers integrating the NVIDIA HGX Vera Rubin NVL8, to accommodate various AI factory and data center deployments.

At COMPUTEX 2026, ASRock Rack showcased AI server platforms integrating advanced liquid cooling solutions from ecosystem partners. Exhibits included the NVIDIA Vera Rubin NVL72 cooled by a dual-rack-width liquid-to-air coolant distribution unit (CDU), and a liquid-cooled system integrating the NVIDIA HGX Rubin NVL8. The fully liquid-cooled 2U16X-GNR2/DLC and 5U16X-GNR2/DLC systems combine CPU and GPU liquid cooling. Additionally, the company demonstrated rack-level liquid cooling solutions using in-row CDUs, suitable for high-density AI deployments.

ASRock Rack showcased its MGX server product line for enterprise and edge environments, including the 6UXGM-GNR2/DLC system supporting up to eight liquid-cooled NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, and the 4UXGM-GNR2 CX8 NVIDIA RTX PRO server optimized for inference cloud environments. For the industrial edge, the 2UXGI-Thor system based on the NVIDIA IGX Thor platform demonstrated physical AI capabilities, enabling real-time sensor processing and functional safety for next-generation autonomous robotics and medical applications.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com