NTS AI Software Stacks

Purpose-Built for Every Workflow – From Instant Productivity to Full Customization

MITAC 4U G4520G6U2BC-N

4U2S 8-GPU AI Server with 8 NVMe U.2 Support

32 RDIMM slots with DDR5-6400 (1DPC) / 5200MHz (2DPC) / MRDIMM 8000 (1DPC) P core memory support
2 Broadcom PEX89104 PCIe 5.0 switch deployed
11 FH/FL PCIe 5.0 x16
8 NVMe U.2 hot-swap, tool-less drive bays
3 NVMe M.2 22110/2280 slots
3 PCIe 5.0 x16 OCP v3.0 LAN mezzanine slot
8 Hot-swap fans and redundant PSU

4U2S 8-GPU AI Server

The MiTAC G4520G6 is a high-performance 4U dual-socket server engineered for demanding AI, HPC, and data-intensive workloads. Built on the Intel® Xeon® 6 Birchstream platform, it offers exceptional compute, memory, and I/O capabilities, with robust support for GPU acceleration and parallel processing tasks.

Key Platform Highlights

Supports dual Intel® Xeon® 6 Scalable Processors with up to 350W TDP for outstanding multi-core performance
Up to 32 DDR5 DIMM slots (16 per CPU) with 8 memory channels per processor, supporting up to 8,192GB DDR5 RDIMM/3DS RDIMM at 6400 MT/s
Accommodates up to 8 full-height, double-width GPU cards for AI, ML, and rendering tasks
Includes 3 additional PCIe 5.0 x16 expansion slots for accelerators and I/O flexibility
Features 3 PCIe 5.0 x16 OCP v3.0 mezzanine slots supporting LAN speeds up to 400Gb/s
Offers 8 NVMe U.2 tool-less hot-swap bays for high-speed, low-latency storage
Equipped with 8 hot-swap 8038 fans and 4 rear 8080 hot-swap fan modules for optimal cooling
3+1 hot-swap CRPS power supply configuration delivering up to 9,600W with 80+ Titanium efficiency
Integrated IPMI port for enterprise-class remote monitoring and management

Artificial Intelligence (AI) and Machine Learning (ML)

Training Complex Models: The server supports up to 8 full-height, double-width GPU cards, delivering the computational power required for training large-scale deep learning models like natural language processing (NLP) systems, computer vision applications, and recommendation engines.
Inference at Scale: The high CPU-to-GPU bandwidth ensures smooth execution of inference tasks, making it ideal for deploying AI models in production environments.

High-Performance Computing (HPC)

Scientific Research: With dual Intel® Xeon® 6 Scalable Processors and up to 8,192GB of DDR5 memory, the server is an excellent choice for computationally intensive applications in fields such as genomics, molecular dynamics, climate modeling, and astrophysics.
Engineering Simulations: Supports simulation tools used in fields like automotive, aerospace, and energy, allowing organizations to model and analyze complex systems with high precision.

Big Data Analytics

Data Processing and Analysis: The server's support for 8 NVMe U.2 hot-swap bays and PCIe 5.0 lanes enables ultra-fast data ingestion and processing, making it ideal for big data platforms such as Apache Hadoop, Spark, or Presto.
Real-Time Insights: High-speed GPUs accelerate real-time analytics workloads, such as fraud detection, customer behavior analysis, and operational intelligence.

Media and Entertainment

Rendering and Post-Production: The server's GPU density and CPU performance are perfect for rendering complex 3D scenes, visual effects, and animation.
Content Creation and Delivery: Optimized for video encoding, transcoding, and streaming at scale, ensuring smooth delivery of high-quality content to global audiences.

Cloud and Virtualization

Virtual Machine Hosting: With its massive memory capacity and robust CPU performance, the server can host a large number of virtual machines (VMs) for cloud service providers or enterprise private clouds.
Containerized Workloads: Supports Kubernetes and Docker environments, enabling microservices deployment and efficient management of cloud-native applications.

Edge AI and IoT

AI at the Edge: With its GPU capabilities, the server can perform AI inference at the edge for applications like autonomous vehicles, smart factories, and video surveillance.
IoT Gateway: Acts as a central processing hub for IoT sensors and devices, aggregating and analyzing data at the edge to reduce latency and bandwidth costs.

Why Choose the MiTAC G4520G6?

Performance Leadership

Powered by the latest Intel Xeon 6 Scalable Processors and DDR5 memory, it delivers exceptional compute and memory performance.

Optimized for GPUs

Designed to handle up to 8 GPUs with dedicated bandwidth, ensuring seamless execution of parallel workloads.

Scalability and Flexibility

Offers unmatched expandability with PCIe 5.0 slots and tool-less NVMe storage, meeting diverse deployment needs.

Energy Efficiency

Built with 80+ Titanium power supplies and efficient cooling to reduce energy costs without compromising performance.

Reliable and Manageable

Enterprise-grade components and remote management capabilities ensure high uptime and simplified maintenance.

The MiTAC G4520G6 server is a cutting-edge solution for organizations looking to future-proof their infrastructure with a platform designed to tackle the most demanding workloads in AI, HPC, and data analytics.

System	Form Factor 4U Rackmount Chassis Model FT83 Dimension (D x W x H) 32.78" x 17.62" x 6.9" (832.6 x 447.4 x 176mm) Motherboard Name G520G6 # Motherboard Notification # The motherboard not sold separately Gross Weight 56 kg (123.5 lbs) Net weight 41.5 kg (91.5 lbs) Board Dimension Prop. 16.85" x 14.26" (428 x 361.9mm)
Front Panel	Buttons (1) UID / (1) PWR w/ LED / (1) RST LEDs (1) HDD / (1) ID / (1) System Event I/O Ports (2) USB 3.2 Gen.1 ports / (1) COM Port / (1) VGA port
External Drive Bay	Q'ty / Type (8) 2.5" Hot-Swap NVMe SSDs Front Drive Bay Interface (8) NVMe U.2 Front Drive Backplane Support NVMe
System Cooling Configuration	Fan (4) 8080 hot-swap fan modules at rear / (8) 8038 hot-swappable fans
Power Supply	Type CRPS Input Range AC 200-240V/16A Frequency 50/60 Hz Output Watts 9,600Watts Efficiency 80 plus Titanium Redundancy 3+1
Processor	Q'ty / Socket Type (2) LGA4710 Supported CPU Series Xeon® 6 Processor Thermal Design Power Wattage Max up to 350W* (TDP)
Expansion Slots	PCIe (11) PCIe Gen 5 x16 slots (FH/FL) Others (3) PCIe Gen.5 x16 OCP 3.0 mezzanine slots
LAN	Q'ty / Port (1) GbE dedicated for IPMI PHY Realtek RTL8211F
Storage NVMe	Connector (M.2) (3) 22110/2280 (by PCIe Gen 4 interface)
Graphic	Connector type D-Sub 15-pin Resolution Up to 1920x1200 Chipset Aspeed AST2600
I/O Ports	USB (2) USB3.0 ports / (2) USB3.2 Gen.1 ports COM (1) DB-9 COM port VGA (1) D-Sub 15-pin port RJ-45 (1) GbE port dedicated for IPMI
TPM (Optional)	TPM Support Please refer to our TPM supported list. Interface SPI
System Monitoring	Chipset Aspeed AST2600 Temperature Monitors temperature for CPU & memory & system environment Voltage Monitors voltage for CPU, memory, chipset & power supply LED Over temperature warning indicator / Fan & PSU fail LED indicator
Server Management	Onboard Chipset Onboard Aspeed AST2600 AST2600 iKVM Feature 24-bit high quality video compression / Supports storage over IP and remote platform-flash / USB 2.0 virtual hub AST2600 IPMI Feature IPMI 2.0 compliant baseboard management controller (BMC) / 10/100/1000 Mb/s MAC interface
BIOS	Brand / ROM size AMI / 64MB Feature Hardware Monitor / FAN speed control automatic / Boot from USB device/PXE via LAN/Storage / Console Redirection / SMBIOS 3.0/PnP/Wake on LAN / ACPI 6.1 / ACPI sleeping states S0,S1,S5
Operating Environment	Operating Temp. 10° C ~ 35° C (50° F~ 95° F) Non-operating Temp. - 40° C ~ 70° C (-40° F ~ 158° F) In/Non-operating Humidity 90%, non-condensing at 35° C
Regulation	FCC (SDoC) Class A CE (DoC) Class A CB Yes RCM Yes VCCI Yes BSMI Yes NRTL Yes
RoHS	RoHSRoHS 6/6 Compliant Yes RoHS 2.0 Yes
Package Contains	Package Contains Barebone (1) G4520G6 Barebone Manual (1) Quick Installation Guide Others (2) E2A Carrier / (2) E2B Carrier

OPERATING SYSTEM

No Operating System
Rocky Linux 8 [+$49.00]
Rocky Linux 9 [+$49.00]
Ubuntu 24.04 LTS [+$49.00]

NTS AI Stack See what's inside each NTS AI package

No NTS AI Stack
NTS AI Stack - LaunchPad
NTS AI Stack - FlexBox
NTS AI Stack - ForgeKit
Other Software Requests - If you require specific software such as AMBER, RELION for cryo-EM, Schrödinger, NVIDIA AI Enterprise, or any other specialized tools, please include your request in the notes section. Our engineering team will assist with optimization and provide a tailored quote.
Please provide details:

Call for pricing