LP_INNO3D_Enterprise_4U_AIR-CX8_html.jpg LP_INNO3D_Enterprise_4U_AIR-CX8_html.jpg
AGS-4UMGX-R1
Accelerate your industrial integration with Artificial Intelligence

INNO3D NVIDIA MGX™ 4U GPU SERVER WITH NVIDIA CONNECTX® 8 SUPERNIC™ - AGS-4UMGX-R1

INNO3D NVIDIA MGX™ 4U GPU SERVER WITH NVIDIA CONNECTX® 8 SUPERNIC™ - AGS-4UMGX-R1
KEY FEATURES
SPECIFICATION
PRODUCT IMAGES
LP_INNO3D_Enterprise_4U_AIR-CX8_html_bg.jpg
LP_INNO3D_Enterprise_4U_AIR-CX8_html_1820x1024.png
INNO3D MGX 4U GPU Server SKU.

INNO3D NVIDIA MGX™ 4U GPU SERVER WITH NVIDIA CONNECTX® 8 SUPERNIC™

AGS-4UMGX-R1

feature icon combo

INNO3D Servers harness the parallel processing prowess of GPUs to empower enterprises in enhancing their capabilities across High-Performance Computing (HPC), data analytics, genomics, 3D modelling and imaging, machine learning, simulations, deep learning, AI, and other mission-critical requirements. These high-performance servers are tailored to excel in specialized AI training, supervised fine-tuning, AI inference, and beyond, offering augmented performance for diverse operational needs.

icon01

LP_INNO3D_Enterprise_4U_AIR-CX8_html_enterprise.png
INNO3D NVIDIA MGX™ 4U GPU SERVER WITH NVIDIA CONNECTX® 8 SUPERNIC™ 
CHASSIS POWERED BY CHENBRO
 

INNO3D MGX 4U GPU Server SKU FEATURE HIGHLIGHTS:

SKU: AGS-4UMGX-R1

  • 4U Rackmount NVIDIA® MGX form factor GPU Server 
  • 2 x Intel® Xeon 6th GEN Scalable processor support
  • 8 Channels DDR5 RDIMM per CPU / Max. 32 RDIMMs 
  • 8 x PCIe Gen 5 x 16 Slots for NVIDIA® RTX PRO 6000 Blackwell Server Edition GPU card support
  • NVIDIA® ConnextX-8 SuperNIC – 8 x 400Gbps QSFP ports
  • 2 x 10GbE Ethernet ports, and 1GbE Management port (BMC)
  • 8 x 2.5 inch front Hot-Swap storage drive bay, up to 8 x E1.S form factor drive (25mm)
  • On-board TPM 2.0 hardware root of trust

  • 4 x 3200W (3+1 redundant) CRPS 80 Plus Titanium Power supply units

    icon-intel

Applications:

  • NVIDIA Omniverse Enterprise Server
  • Large scale Digital Twins
  • Accelerated computing / HPC
  • LLM inference & Fine tuning

 

Additional Features

UP TO 10 GPU READY

Offering high-performance computing capabilities for a wide range of demanding applications.
*GPU sold separately. Up to 10 GPU dual-slot GPUs ready

MODULAR DESIGN

Installing or upgrading components and maintaining server products are more easily accomplished with a streamlined modular design for easier access to internal hardware.
*Available on select models

HIGH SPACE TO PERFORMANCE RATIO

A smartly designed chassis provides more space when compared to those of similar size and form factor, providing up to 10 full-length full-height, dual-slot graphics cards to offer a greater space to performance advantage.
*select models

PCIe DIRECT DRIVE

Enabling flexible Server use scenarios on a wide scale: 
Storage Focus
Prioritize PCIe connections to storage
Compute Focus
Prioritize more PCIe bandwidth for maximum GPU capabilities
Multi-server Focus
Enable more PCIe bandwidth towards high-speed network connectivity between Servers for even more computational performance
 

POWERFUL COOLING

Equipped to cool with powerful air coolers or high-performance liquid-cooled hardware, the streamlined design of INNO3D GPU Servers maximizes efficient heat dissipation to provide greater cooling to the entire machine, especially critical components like the CPU and GPU to enable higher and longer sustained performance.
*The number and performance equipped fans may vary by model

REDUNDANT POWER

With support with up to 4+1 2000W 80 PLUS Titanium redundant power supplies, INNO3D GPU Servers can further maximize efficiency to provide uninterrupted operation, reduce operating costs, and enable easier servicing.

DUAL ROM
ARCHITECTURE

If the ROM that stores the BMC and BIOS fails on INNO3D GPU Servers, the system will automatically reboot with the backup for uninterrupted operation. Once the primary BMC is updated, the ROM of the backup BMC will automatically update the backup through synchronization.
*Available on select models

TRUSTED PLATFORM
MODULE 2.0 READY

INNO3D GPU Servers include a TPM header to support an optional TPM 2.0 module for hardware-based authentication and sensitive data storage.
*Available on select models

LP_2024_INNO3D_GPU_server_side_banner-01.jpg

FROM HIGH-PERFORMANCE
COMPUTING WORKSTATIONS
TO ENTERPRISE SERVER

Handle standard small & medium workloads with the right-sized performance in a Server grade Tower Workstation or take on business-critical demands with scalable rack-mounted GPU 6U Servers.
 

 inno3d_logo

nvidia omniverse

3D design collaboration and simulation and more, NVIDIA Omniverse™ is a computing platform that enables individuals and teams to develop Universal Scene Description (OpenUSD)-based 3D workflows and applications. From concept design for animated movies, to industrial digital twins of factories, Omniverse transforms the way teams approach building and operating metaverse applications.

NVIDIA AI Enterprise

NVIDIA AI Enterprise is a cloud-native suite of software tools, libraries, and frameworks, including NVIDIA NIM™ and NeMo™ microservices, that accelerate and simplify the development, deployment, and scaling of AI applications.

NVIDIA Omniverse

NVIDIA Omniverse is a platform of APIs, SDKs, and services that enable developers to integrate OpenUSD, NVIDIA RTX rendering technologies, and generative physical AI into existing software tools and simulation workflows for industrial and robotic use cases.

NVIDIA AI Factory Validated Design

The NVIDIA Enterprise AI Factory validated design provides guidance for developing, deploying, and managing agentic AI, physical AI, and HPC workloads on NVIDIA RTX PRO servers. Designed for enterprise IT, it recommends accelerated computing, networking, storage, and software to help deliver faster time-to-value AI factory deployments while mitigating deployment risks.

NVIDIA AI Data Platform

RTX PRO Servers are a key building block of the NVIDIA AI Data Platform- a customizable reference design for a new class of AI infrastructure—one that integrates enterprise storage with NVIDIA-accelerated computing to power AI agents with near-real-time business insights.

NVIDIA Run:ai

Run:ai is an AI workload and GPU orchestration platform that helps enterprises maximize RTX PRO Servers. By increasing GPU utilization and overall throughput, it accelerates AI workloads, minimizes bottlenecks, and makes development faster and more efficient.

LP_INNO3D_Enterprise_Tower_AWS-511Z-R1.jpg

inno3d x nvidia

 

NVIDIA RTX PRO 6000
Blackwell Workstation Edition

Powering the next era of AI.

 

DATASHEET

 

nvidia-omniverse-creator-bb580_440-d.jpg LP_2024_INNO3D_GPU_server_side_banner_mobile01.jpg

 

CREATE IN 3D FASTER
THAN EVER

Sync your favorite creative apps to Omniverse and USD and work with your 3D data in a unified view.

nvidia-omniverse-developer-bb580_440-d.jpg LP_2024_INNO3D_GPU_server_side_banner_mobile02.jpg

AUGMENTED BY AI

Build custom extensions, tools, and microservices to
accelerate your 3D workflows, generate synthetic data,
and build industrial metaverse applications.

nvidia-omniverse-enterprise-bb580_440-d.jpg LP_2024_INNO3D_GPU_server_side_banner_mobile03.jpg

 

 

POWER INDUSTRIAL
DIGITALIZATION

Break 3D data silos, unify large enterprise teams, and
enable the next wave of 3D and simulation workflows.

INTEL® XEON® SCALABLE PROCESSORS

Designed to accelerate performance across fast-growing workloads, Intel Xeon Scalable Processors have powerful built-in accelerators to help improve performance efficiency for emerging workloads like AI and deep learning applications. These built-in accelerators can also ensure maximum efficiency in the core and power utilization, optimizing operational sustainability.

intel-logo

Intel® Advanced Matrix Extensions
(Intel® AMX)

To boost AI inference and training with 2D tiled operation

Intel® QuickAssist Technology

For faster encryption and bulk data compression

Intel® In-Memory Analytics Accelerator
(Intel® IAA)

To boost database performance with lower memory bandwidth

Intel® Data Streaming Accelerator
(Intel® DSA)

To speed-up data movement

Intel® Dynamic Load Balancer
(Intel® DLB)

To optimize queue scheduling and packet processing

 

 

 

LP_20240530_RTX_AI_PC_Campaign_GTMK_white_bg.jpg

AMD EPYC™ PROCESSORS POWERING THE NEXT GENERATION OF SERVER ARCHITECTURE AND ENERGY EFFICIENCY

The new family of EPYC processors deliver even better CPU performance and performance per watt and do so on a platform with 2x the throughput of PCIe 4.0 lanes that also has support for 50% more memory channels. For this new platform, ZOTAC 4th Gen AMD EPYC supported Servers can get the most out of EPYC-based systems that support fast PCIe Gen5 accelerators and Gen5 NVMe drives, in addition to support for high performant DDR5 memory.

EPYC-logo

4th Gen AMD EPYC Processors for SP5 Socket

 

5nm architecture

Compute density increased with more transistors packed in less space

128 CPU cores

Dedicated cores and targeted workloads for Zen 4c & Zen 4 cores

Large L3 cache

Select CPUs have 3x or more L3 cache for technical computing

SP5 compatibility

All 9004 series processors are supported on one platform

DDR5 memory

Increased memory throughput and higher DDR5 capacity per DIMM

PCIe 5.0 lanes

Increased IO throughput achieving 128GB/s bandwidth in PCIe x16 lanes

 

 inno3d_logo

Privacy Policy | Contact Us

 

DISCLAIMER

The information presented in this page is for informational purposes only and may contain technical inaccuracies, omissions and typographical errors.  The information contained herein is subject to change and may be rendered inaccurate for many reasons, including but not limited to product and specification changes, component and motherboard version changes, new model and/or product releases, product difference between differing manufacturers, software changes, BIOS flashes, firmware upgrades, or this like.  InnoVISION assumes no obligation to update or otherwise correct or revise this information.  However, InnoVISION reserves the rights to revise this information and to make changes from time to time to the content hereof without obligation of InnoVISION to notify any person of such revision or changes.
InnoVISION makes no representations or warranties with respect to the contents hereof and assumes no responsibility for any inaccuracies, errors or omissions that may appear in this information.

 

Copyright Notice

© Copyright 2026, InnoVISION Multimedia Limited.  All rights reserved.  The INNO3D®, INNO3D Brutal by Nature Logo, are the registered and/or trademarks of InnoVISION Multimedia Limited in Asia, Europe and / or other countries.   NVIDIA®, NVIDIA logo are registered and/or trademarks of NVIDIA Corporation in the United States and other countries.  All other trademarks and registered trademarks are the property of their respective owners.

 

Datasheet:
Datasheet
PDF
Download Data Sheet
Computing System
Computing System
Reference Number

INNO3D MGX® 4U GPU Server /ConnextX-8 SuperNIC
AGS-4UMGX-R1

Form Factor

NVIDIA® MGX 4U Rackmount

CPU

Intel® Granite Rapids / Sierra Forest
Intel® Xeon® 6 6500 / 6700 processor support

Socket

2 x LGA4710 (Package Carrier E2A / E2B)

Chipset

N/A

Memory

8-channel (2DPC) DDR5 RDIMM / MRDIMM per CPUMEMORY
DDR5 6400 - 1DPC / DDR5 5200 - 2DPC
32 x DIMMs total
Memory modules: 32GB, 64GB, 96GB, 128GB

Graphics:
Graphics
Video

Integrated Aspeed® AST2600
1920x1200@60Hz, DDR4 SDRAM

Storage:
Storage
Storage

Internal:
2 x NVMe PCIe Gen 5 x4 M.2 Slot (M-Key 22110 / 2280) 

Drive Bay:
8 x 2.5 inch Front Hot-Swap Storage Bays (up to 8 x E1.S 25mm height)

RAID

Intel® SATA RAID 0/1/10/5

I/O Ports:
I/O Ports
LAN

8 x 400Gbps ports (QSFP, NVIDIA® ConnextX®-8 SuperNIC)
2 x 10GbE LAN ports
1 x 1GbE management port (BMC)

Expansion Slots

- 8 x PCIe Gen 5 x 16 Slot (x16 PCIe Signal) - Configured with 8 x NVIDIA RTX PRO™ 6000 Blackwell Server Edition GPUs
- 1 x PCIe Gen 5 x 16 Slot (x16 PCIe Signal) - Configured with 1 x NVIDIA® BlueField®-3 DPU

Front I/O

8 x 400Gbs QSFP ports
2 x USB 3.2 Gen 1 Type A ports
2 x 10GbE Ethernet port
1 x 1GbE BMC management port
1 x Mini-DisplayPort

Mechanical:
Mechanical
Power Supply

4 x Redundant 3200W CRPS 80 Plus Titanium PSU (3+1)

Cooling

Air Cool:
5 x internal 80x80x80 mm Dual Rotar Ball Bearing Fan for GPU Zone Air cooling
5 x internal 80x80x80 mm Dual Rotar Ball Bearing Fan for HPM Zone Air cooling
(Hot Swapping support)

Operation Properties

Operating temperature: 0°C to 35°C
Operating humidity: 8-90% (non-condensing)
Non-operating temperature: -20°C to 70°C
Non-operating humidity: 5%-85% (non-condensing)

Product Dimensions

85.8cm (Depth) x 44cm (Width) x 17.5cm (Height)

Miscellaneous:
Miscellaneous
OS Compatibility

Windows Server 2022 / 2019  (x64)
Red Hat Enterprise Linux 9.1 / 9.0/ 8.7 (x64) 
Ubuntu 22.04 LTS (x64) 
VMware ESXi 8.0 / 7.0 Update2 
Citrix Hypervisor 8.2.0 

TPM

On-board TPM 2.0
Hardware Root of Trust - Microchip CEC1736 

Contents

1 x AGS-4UMGX-R1 4U Rackmount System
2 x CPU heatsinks
4 x CPU Package Carriers
1 x L-shape Rail Kit
4 x CRPS (Pre-installed)
10 x System Cooling fans (Pre-installed)
CPU and Memory (SKU option)
Internal Storage device (SKU option)
Front Panel Storage device (SKU option)

Available SKU:
Available SKU
LP_INNO3D_Enterprise_4U_AIR-CX8_html_enterprise.png
LP_INNO3D_Enterprise_4U_AIR-CX8_html_enterprise.png
TOP