ASUS has announced its participation in the 2025 OCP Global Summit, where loads of new AI servers and solutions are being showcased there.
One of the centerpieces from ASUS is dedicated to its AI Factory solution, an end-to-end framework built on NVIDIA’s Blackwell architecture. The ecosystem features ASUS AI PODs powered by the GB300 NVL72 and XA NB3I-E12 servers accelerated by NVIDIA HGX B300 systems, serving as foundational components for enterprise AI factories.
It combines cutting-edge hardware, optimized software, and professional services to accelerate AI adoption from edge to cloud. It supports diverse use cases from generative AI and natural language processing to predictive analytics, while simplifying deployment and maximizing resource efficiency. By combining ASUS servers, rack-scale AI PODs, and high-serviceability designs, the platform streamlines operations, reduces complexity, and scales seamlessly to meet the demands of modern AI development.
In addition, there are server solutions powered by AMD EPYC 9005 processors, offering high density and performance for AI-driven and mission-critical data center workloads. For example, the ESC8000A-E13X accelerates generative AI and LLM applications with support for NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs and NVIDIA ConnectX-8 SuperNICs, providing 400G InfiniBand or Ethernet connectivity for ultra-low latency and massive bandwidth.
Meanwhile, the RS520QA-E13 series delivers high-performance multi-node computing for HPC, EDA, and cloud workloads, featuring up to 20 DIMM slots per node, advanced CXL memory expansion, PCIe 5.0, and OCP 3.0 for next-level efficiency.
There’s also the Ascent GX10, which has been previewed several months before and is now available globally, and with this NVIDIA GB10 Superchip, 1 petaFLOPs of AI performance capable of running LLM of up to 200B parameters are easily obtainable and sit within one’s palm. The machine also comes with 128GB of unified memory for a streamlined development cycle, sitting nicely on a table.
