Asus Unveils XA P2N-E2, an Nvidia Vera CPU-Based Server for Agentic AI
TECHWORLD ·
✦ AI Summary
Asus unveiled the server "XA P2N-E2" for agentic AI and reinforcement learning workloads at the "OCP Global Summit 2026."
The "XA P2N-E2" is based on the Nvidia MGX architecture and the Nvidia Vera CPU, and supports up to two Nvidia RTX Pro 6000 Blackwell Server Edition GPUs.
In addition to the XA P2N-E2, Asus also unveiled the XA VR721-E3, the XR R1I12B-M1BHFS, a product lineup based on the Nvidia Vera Rubin platform, and server platforms based on Intel and AMD processors.
Asus participated in the "OCP Global Summit 2026" and unveiled its latest server, the "XA P2N-E2," optimized for agentic AI and reinforcement learning workloads. The product is built on a 2U form factor.
The "XA P2N-E2" is designed based on the Nvidia MGX architecture and is equipped with the Nvidia Vera CPU. It also supports up to two Nvidia RTX Pro 6000 Blackwell Server Edition GPUs and is configured as a compact 2U platform that combines high-performance CPU computing with flexible GPU acceleration.
Asus said the "XA P2N-E2" is designed to efficiently handle code execution, sandboxing, and data processing. It also said the product can handle not only AI model inference but also the core execution layer of applications.
Asus explained that high single-thread performance helps maintain AI agent responsiveness. It also said high-bandwidth, low-power memory helps maintain AI agent responsiveness during complex execution and evaluation tasks.
Asus unveiled a full-stack AI infrastructure solution at the OCP Global Summit. The scope of the unveiling covered XA P2N-E2, rack-scale AI systems, accelerated computing, high-speed networking, context memory, ultra-low-latency inference, and high-density storage.
Asus showcased the "XA VR721-E3" as a flagship system. The "XA VR721-E3" integrates the Nvidia Vera Rubin NVL72 platform and is built with 36 Nvidia Vera processors and 72 Nvidia Rubin GPUs. It also applies a fully liquid-cooled architecture for large-scale AI training and inference, and offers high cooling efficiency.
Asus also unveiled the "XR R1I12B-M1BHFS." The "XR R1I12B-M1BHFS" applies direct liquid cooling to the GPU baseboard chip based on Nvidia HGX Rubin NVL8, while air cooling is applied to the Intel Xeon 6 processors.
Asus then unveiled additional product lines based on the Nvidia Vera Rubin platform. The purpose of these product lines is to implement essential elements of a full-stack AI factory, including high-speed networking, scalable context memory, and a fast inference environment.
The company first unveiled a next-generation AI infrastructure that combines four core technologies. The core technologies are Nvidia HGX Rubin NVL8, Nvidia Spectrum-6 Ethernet switches for ultra-fast AI networking, Nvidia CMX for scalable context memory storage, and Nvidia Groq 3 LPX, an inference accelerator for token generation in agentic systems.
It also unveiled a large lineup of server platforms based on Intel and AMD processors to address customer needs across cloud, AI, and HPC. With these offerings, the company presented a combined configuration to meet demand across cloud, AI, and HPC.
Among them, the exhibits included the 1U dual-socket platform "RS700-E12-RS4U" with Intel Xeon 6 processors and the 2U dual-socket platform "RS720-E12-RS12U" with Intel Xeon 6 processors. Both platforms share the DC-MHS architecture and Tool-less maintenance design, and are configured to provide modular deployment flexibility and flexibility for data-intensive workloads.
For the AMD-based lineup, a server platform and an AI acceleration system were introduced together. In the server segment, the 1U dual-socket platform RS700A-E14B-R16U equipped with the 6th-generation AMD EPYC 9006 series, code-named Venice, was on display.
The RS700A-E14B-R16U supports the PCIe 6.0 interface and supports up to 16 E3.S drives. It also improves thermal management efficiency by applying Asus Thermal Radar 3.0.
In the AI acceleration segment, the ESC8000A-E13P supporting AMD Instinct MI350P and the 10U form factor XA AM3A-E13 were shown together. The XA AM3A-E13 integrates eight AMD Instinct MI355X accelerators and emphasizes a configuration suited for high-density AI environments.
Source: TECHWORLD · Lee Gwang-jae
Original: https://www.epnc.co.kr/news/articleView.html?idxno=409096
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.