ASUS Unveils XA P2N-E2 Powered by NVIDIA Vera CPU for Agentic AI Workloads at OCP 2026

ASUS unveiled the XA P2N-E2, a 2U server based on NVIDIA MGX architecture and powered by NVIDIA Vera CPUs, at OCP Global Summit 2026. The system targets agentic AI and reinforcement learning workloads, including code execution, tool use, and data processing beyond the AI model itself. It supports up to two dual-slot NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, combining high-performance CPU computing with GPU acceleration in a compact 2U form factor. The server emphasizes single-thread CPU performance and energy-efficient memory to maintain responsiveness during demanding AI tasks.

ASUS also showcased the XA VR721-E3 AI POD, integrating 36 NVIDIA Vera CPUs and 72 NVIDIA Rubin GPUs via the NVIDIA Vera Rubin NVL72 platform. This rack-scale system uses a fully liquid-cooled architecture for large-scale AI training, post-training, and inference. Additionally, the XR R1I12B-M1BHFS combines NVIDIA HGX Rubin NVL8 with direct-to-chip liquid cooling for GPUs and air cooling for Intel Xeon 6 processors, offering flexible server configurations for varying deployment needs.

The ASUS booth featured four NVIDIA technologies: NVIDIA HGX Rubin NVL8 for accelerated computing, NVIDIA Spectrum-X Spectrum-6 Ethernet for high-speed AI networking, NVIDIA CMX Context Memory Storage Platform for scalable context memory, and NVIDIA Groq 3 LPX for ultrafast inference acceleration. These components are designed to optimize AI factory performance, enabling low-latency and large-context requirements for agentic systems. The full-stack architecture supports scaling from individual servers to rack-scale deployments while improving performance per watt.

ASUS highlighted additional platforms at the event, including Intel-based RS700-E12-RS4U and RS720-E12-RS12U servers powered by Intel Xeon 6 processors, and the AMD-based RS700A-E14B-R16U with 6th Gen AMD EPYC 9006-series processors. The company also demonstrated systems like the ESC8000A-E13P with AMD Instinct MI350P accelerators and the XA AM3A-E13 with eight AMD Instinct MI355X accelerators, emphasizing air-cooled, high-density AI deployments without dedicated chilled-water infrastructure.