#USA #San Jose #NVIDIA #ASUS #XA P2N-E2

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

At the OCP Global Summit 2026, ASUS revealed its latest addition to the AI infrastructure market, the XA P2N-E2. This 2U server, built on NVIDIA's MGX architecture, harnesses the power of NVIDIA Vera CPUs to tackle demanding agentic AI workloads, including reinforcement learning. This launch highlights ASUS's commitment to advancing AI systems designed not only for executing models but also for handling extensive application processes like code execution and data orchestration.

Key Features of the XA P2N-E2

The XA P2N-E2 stands out with its ability to support up to two dual-slot NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. This combination of high-performance CPU computing and flexible GPU acceleration enables efficient handling of complex AI tasks while maintaining a compact dual-unit form factor. With a focus on responsive performance, the server utilizes high-bandwidth, energy-efficient memory designed to ensure AI agents remain agile during their most intense operations.

In addition to its specialized hardware, ASUS presented its broader AI portfolio at the summit, ranging from rack-scale systems to high-speed networking and low-latency storage. Notably, the XA VR721-E3, part of their AI POD solution, integrates 36 NVIDIA Vera CPUs and 72 NVIDIA Rubin GPUs, showcasing a fully liquid-cooled architecture for extensive AI training and inference tasks.

Advancing AI Technology with Modular Solutions

ASUS’s commitment extends beyond mere hardware; the firm has crafted modular and scalable AI infrastructures that adapt to varying workloads. Their NVIDIA-accelerated systems facilitate diverse deployment models, optimizing compute density and enhancing performance for every unique requirement.

Moreover, with NVIDIA technologies such as HGX Rubin NVL8 for computing acceleration and NVIDIA Spectrum-X for high-speed networking, ASUS is able to ensure that users can efficiently manage data movement and workload performance within complex AI frameworks.

A focal point of their showcase was the integration of NVIDIA’s DSX reference design, which allows for the creation of high-fidelity digital twins—technological representations that optimize factory designs before deployment, consequently streamlining operations.

Future-Ready AI Factories

ASUS is not only building powerful servers but is also pioneering the creation of what are termed AI factories. These environments are envisaged to leverage robust architectures for a holistic deployment of AI capabilities across various industries. The XA P2N-E2 serves as a linchpin in this endeavor, providing the essential performance needed for today’s demanding AI applications.

The utility of NVIDIA's Groq 3 LPX system accentuates these efforts, providing ultra-fast token generation for enhanced inference responsiveness. This is a critical component as agentic AI systems thrive on rapid processing capabilities.

The ASUS Ecosystem for AI Workloads

ASUS further diversifies its offerings with the incorporation of Intel Xeon 6 processors and AMD EPYC 9006 series processors across its platforms. This flexibility allows customers to choose architectures that meet a spectrum of needs from AI and cloud computing to high-performance computing.

For enterprises requiring resilient solutions, their 1U dual-socket platform, the RS700A-E14B-R16U, blends robust architecture with leading-edge connectivity, ensuring efficiency and performance even under heavy operational demands.

In summary, ASUS's unveiling of the XA P2N-E2 marks a significant leap forward in the AI server landscape. As organizations strive to advance their AI initiatives, the innovative features and scalable solutions offered by ASUS will empower users to meet the future needs of technology-driven industries.

Visitors to the OCP Global Summit can explore these advancements firsthand at ASUS Booth #B9 and learn more about how these technologies can transform their AI infrastructures.