KEY POINTS
- Accelerating token generation: ASUS delivers rack-scale AI infrastructure and deployment capabilities designed to reduce time to first token
- Strengthening context memory: High-performance storage and data architectures optimized for scalable AI inference and more
- Enabling enterprise AI adoption: Integrated multi-turn and agentic AI workflows supporting secure operations and innovations for diverse industries
TAIPEI, Taiwan — ASUS today announced its latest AI infrastructure solutions showcase at Computex 2026, delivering end-to-end AI factory capabilities in collaboration with ecosystem leaders including NVIDIA, Intel® and AMD. Spanning rack-scale AI factories and POD architectures to ultrafast content memory storage solutions and enterprise-ready agentic AI applications, ASUS provides enterprise with a complete foundation for the entire AI lifecycle — from infrastructure design and deployment to large-scale token generation, inference, and real-world business adoption.
By providing flexible, tightly integrated solutions across compute, storage, networking, and software, ASUS empowers enterprises to build, deploy, and optimize high-performance AI infrastructure at scale —significantly accelerating time to market while maximizing operational efficiency, cost-effectiveness, and strategic agility in an increasingly AI-driven competitive landscape.
Accelerating token generation with NVIDIA DSX
ASUS is at the forefront of AI infrastructure with the XA VR721-E3, ASUS AI POD built on NVIDIA Vera Rubin NVL72, a 100% liquid-cooled rack-scale platform purpose-built for trillion-parameter models and next-generation AI factories. To facilitate the deployment of these advanced AI capabilities and leverage the NVIDIA DSX platform for token generation, ASUS assists customers to simulate the AI factory blueprints into deployment-ready infrastructure. By integrating ASUS rack-scale AI systems with NVIDIA DSX customers can comprehensively evaluate critical deployment factors — including power delivery, cooling design, networking topology, storage integration, and facility readiness — before physical buildout commences. This is further enhanced through OpenUSD-based digital-twin workflows, where ASUS and ecosystem partner solutions are represented as simulation-ready assets, enabling customers to accurately anticipate infrastructure requirements and optimize deployment plans in advance.
Addressing rigorous AI workload demands, ASUS will also showcase systems accelerated by NVIDIA HGX Rubin NVL8 and Intel® Xeon® 6 processor, including XA NR1I-E12L, an innovative hybrid-cooled option, and XA NR1I-E12LR, a 100% liquid-cooled system. The lineup also includes XA NB3I-E12, featuring NVIDIA HGX B300, delivering high-performance acceleration for training, inference and simulation. The portfolio is further strengthened by a comprehensive lineup of high-performance, scalable NVIDIA MGX servers. Making their debut are XA P8A-E14AXL, a high-density 6U system supporting eight liquid-cooled NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs; and XA P4N-E2, a 2U AI server powered by NVIDIA Vera C2 processors for next-generation agentic AI systems. ASUS will also showcase its ESC8000 series, including ESC8000-E12P and ESC8000A-E13P, supporting NVIDIA RTX PRO 6000 Blackwell and RTX PRO 4500 Blackwell Server Edition GPUs for demanding AI, data processing, video and visual computing workloads, as well as ESC8000A-E13X, integrated with NVIDIA ConnectX-8 SuperNICs for extreme GPU-to-GPU connectivity.
