Out of stock

$8,699^.00

Free 90-day Returns

Sold by Newegg

Shipped by Newegg

Price alert

Add to Wish List

Main image of ASUS ESC8000A-E13-32W 4U 8-GPU AI Server | Dual AMD EPYC 9005 SP5 | 24x DDR5 ECC | 8x Hot-Swap NVMe | PCIe 5.0 | NVLink | 3+1 Redundant 3200W Titanium PSU

Shop All ASUS Products

ASUS ESC8000A-E13-32W 4U 8-GPU AI Server | Dual AMD EPYC 9005 SP5 | 24x DDR5 ECC | 8x Hot-Swap NVMe | PCIe 5.0 | NVLink | 3+1 Redundant 3200W Titanium PSU

Be the first to review this product...

"4U flagship AMD GPU chassis, supports up to 8 dual-slot full-size NVIDIA H200 / RTX PRO 6000 AI accelerators
Dual AMD SP5 socket, EPYC 9005 series up to 192 cores, max 500W CPU TDP
24x DDR5 ECC registered memory, 12-channel architecture, max 3TB capacity, up to 6400MHz
8 hot-swappable NVMe all-flash high-speed storage bays
Full PCIe Gen 5.0 expansion lanes, NVIDIA NVLink GPU interconnection support
ASMB11-iKVM remote BMC out-of-band management, SEV-SNP & TPM-SPI enterprise security
3+1 redundant 3200W Titanium ultra-stable mission-critical power supply"

Overview

Reviews

See all Q & A

Any questions? Ask Celeste^SM will help you find out quickly.

Try me now

ESC8000A-E13

AMD EPYC™ 9005 dual-processor 4U server that supports eight dual-slot NVIDIA® H200 GPUs or NVIDIA RTX PRO™ 6000 Blackwell Server Edition, NVIDIA RTX PRO™ 4500 Blackwell Server Edition, up to 24 DIMM, two PCIe®5.0 slots and up to six U.2 NVMe drives.

Form Factor

CPU Number

Memory Number

KEY FEATURES

High-performance server configuration: Equipped with AMD EPYC 9005 processors featuring 192 Zen 5c cores, 12-channel DDR5 memory up to 6400MHz and a maximum TDP of 500 watts per socket.
Direct GPU support: Designed as a high-density 4U server capable of housing eight dual-slot NVIDIA H200 GPUs or NVIDIA RTX PRO™ 6000 Blackwell Server Edition, NVIDIA RTX PRO™ 4500 Blackwell Server Edition,each supporting up to 600 watts.
Optimized connectivity: Offers two PCIe 5.0 slots for high-bandwidth NICs and DPUs, connected directly to CPU to reduce latency.
Storage flexibility: Includes options for hardware RAID function support, configurations with additional support for up to six 2.5” U.2 NVMe drives
Efficient maintenance: Features ASUS-exclusive toolless design for streamlined maintenance and maximum efficiency.
Advanced management: Includes ASUS Control Center IT-management software and a hardware-level Root-of-Trust solution for enhanced security and control.

Eight dual-slot GPUs in a high-density 4U compute server

ASUS ESC8000A-E13, powered by AMD EPYC™ 9005 processors, delivers outstanding performance with top-tier GPUs, faster GPU interconnects, and a high-bandwidth fabric. This robust server directly supports up to eight dual-slot GPUs, available in both active and passive configurations, and provides flexible scalability through options like the NVIDIA® NVLink® Bridge. This setup ensures seamless scaling and performance optimization, making ESC8000A-E13 a perfect solution for handling demanding AI training and inferencing workloads.

Versatile storage capabilities

ESC8000A-E13 offers flexible and scalable storage solutions, supporting up to 24 DIMMs, two PCIe® 5.0 slots, and up to six U.2 NVMe drives with optional RAID/HBA card. This versatile configuration allows organizations to customize their storage infrastructure to meet a wide range of workload requirements, delivering optimal performance and capacity for every need.

System Layout

Panel Layout

Block Diagram

Front Panel

Rear Panel

Unlock next-gen performance with AMD EPYC™ 9005 series

The AMD EPYC™ 9005 series takes AMD's legacy of x86 architecture innovation to new heights, leveraging cutting-edge 3nm technology for record-breaking performance. With the introduction of high-performance DDR5 DIMMs and blazing-fast PCIe 5.0 I/O, EPYC™ 9005 CPUs are engineered to handle the most demanding workloads. Designed to power memory-hungry AI-driven data centers, machine learning (ML), high-performance computing (HPC), and large in-memory computations.

Up to

192 cores

on Zen 5c

128

PCIe 5.0 lanes in a dual-socket server

Optimized server configuration with high-band width PCIe 5.0 and scalable NVMe storage

Delivering next-level performance and an optimized configuration, ESC8000A-E13 is designed to meet the demands of modern workloads. With two direct-connected PCIe 5.0 slots, this server enables seamless integration of high-bandwidth PCIe NICs and DPUs for unparalleled performance scaling. Enhance your storage capabilities with flexible RAID options, supporting up to six U.2 NVMe with an optional HBA/RAID card for data back-up and protection. Whether you're handling data-intensive applications or accelerating compute workloads, ESC8000A-E13 delivers the power and flexibility you need to stay ahead.

Effortless maintenance with a toolless design for speed and simplicity

Designed for easy maintenance with a refined mechanical structure, ESC8000A-E13 allows the entire machine to be disassembled completely without tools – with even the RAID card being easily disengaged from the connector.

Toolless removal

Quick installation

Toolless removal and quick installation for PCIe card

Server-management Solution

BMC Management

ASUS Control Center

TPM 2.0

Remote server management

ASUS ASMB12-iKVM is the latest server-management solution from ASUS, built upon the ASPEED 2600 chipset running on the latest AMI MegaRAC SP-X. The module provides various interfaces to enable out-of-band server management through WebGUI, Intelligent Platform Management Interface (IPMI) and Redfish® API.

Streamline IT operations via a single dashboard

ASUS Control Center (ACC) is remote IT-management software application designed for monitoring hardware and software IT assets and inventory status, and enabling seamless remote BIOS configurations and updates, efficient IT diagnostics and troubleshooting and enhanced security with hotfix updates — allowing easier server management for any IT infrastructure.

Trusted Platform Module 2.0

ASUS servers also include support Trusted Platform Module 2.0 (TPM 2.0) to secure hardware through integrated cryptographic keys and offer regular firmware update for vulnerabilities.

Warranty & Returns

Warranty, Returns, And Additional Information

Warranty

Limited Warranty period (parts): 3 years
Limited Warranty period (labor): 3 years
Read full details

Return Policies

We want to ensure your holiday shopping this year is hassle-free. That’s why we’re automatically extending our return policies on almost every product! For purchases placed between November 1st, 2026 and December 31, 2026 you will have until January 31, 2027 to return your product.

Return for refund until: January 31, 2027
Return for replacement until: January 31, 2027
This item is covered by Newegg.com's Extended Holiday Return Policy.
Read full return policy for details.

Manufacturer Contact Info

The ASUS ESC8000A-E13-32W is a 4U, dual-socket AMD EPYC 9005 server barebones platform designed for up to eight dual-slot NVIDIA GPUs, making it suitable for data centers and enterprise teams building AI training, inference, or HPC systems.

Recommended For

AI teams training models or serving LLM inference across multiple GPUs that need a rack-mounted platform designed to scale to eight accelerators.
HPC and research teams running GPU compute alongside heavily parallel CPU workloads, where dual EPYC processors, high-bandwidth PCIe connectivity, and multiple accelerators serve one shared workload.
Organizations consolidating GPU-accelerated rendering, virtual desktop infrastructure, and AI workloads in a shared rack server rather than buying separate systems for each task.
IT teams operating rack infrastructure that need remote out-of-band management and serviceable GPU and system-fan modules for maintenance.

Real-World Use Scenarios

An AI team trains or fine-tunes models that exceed the capacity of a single accelerator. With a supported multi-GPU configuration and software set up for distributed training, the server can keep the work on one system rather than splitting it across separate machines.
An organization hosts a large language model for internal chat, coding assistance, or other staff tools. A multi-GPU setup can serve models too large for one GPU, while the actual response speed and number of simultaneous users depend on the GPUs, model, and inference software selected.
A team processes large batches of prompts or long-context documents offline—for example, summarizing archives or generating embeddings. With suitable GPUs and inference software, the server can run these jobs locally and handle more work in batches rather than relying only on interactive, one-request-at-a-time processing.
Researchers or engineers run GPU-based high-performance computing jobs, such as compatible scientific or engineering applications. The server consolidates CPU and GPU compute in a rack system; whether a particular application benefits depends on its software support and ability to use multiple GPUs.
An IT team builds a shared AI development or virtual-machine host for multiple projects. It can centralize compute and storage in one server, but user isolation and GPU sharing require compatible virtualization, drivers, and workload management software.

Spec Summary

Attribute	Value
Best For	AI and GPU-accelerated server workloads requiring support for multiple accelerators

Item#: N82E16816110225