From Edge to Trillion-Token AI: ASUS Infrastructure For Every Scale

Author: Jerry Wong, Product Manager, ASUS Collection: 2025 OCP Global Summit - Expo Hall Sessions

Summary

ASUS presents its comprehensive AI infrastructure portfolio spanning edge to data center scale, with over 20 years of hardware experience evolving from board-level to system-level to rack-scale solutions. The product line covers: GB series (Nvidia GB200/GB300 NVL72 AI Pod at 130-140 kW, liquid-cooled), HGX series (B300 with air-cooled and liquid-cooled options), ESC8000 series (MGX 4U compliance supporting 8x Nvidia RTX Pro 6000 via PCIe switch), mentoring/general-purpose servers (RS700 DCMHS compliance for Intel/AMD), and edge devices.

ASUS provides end-to-end data center deployment services including consulting, configuration planning, network/compute/storage topology design, installation, validation, deployment, after-sales maintenance, and cooling infrastructure. The software stack includes AIDC (ASUS Infrastructure Deployment Center for firmware updates and BMC configuration at scale), ACC (ASUS Control Center for IT/cluster/login management), and ASUS AI Hub turnkey application platform with Nvidia software (NVIE).

The company highlights its Taiwan deployment track record: 10+ year partnership with NCHC (National Center for High Performance Computing) providing national-scale computing platforms, and 2024 construction of Taiwan's largest data center with UBLink, achieving Top 500 rank #31 and Green 500 rank #44 using liquid-to-air fan wall infrastructure. The system is powered by Nvidia GB300 with Blackwell Ultra GPUs and NVLink technology.