Skip to product information
1 of 1

Supermicro

Supermicro AS-4125GS-TNRT | H13 4U GPU Server, Dual AMD EPYC 9004/9005, 8x GPU, 24x 2.5" NVMe

SKU:AS-4125GS-TNRT

Regular price POA
POA
POA
Sale Sold out
Taxes included. Shipping calculated at checkout.

Custom Built Servers

Need a different spec, quantity or configuration?

We build configured-to-order refurbished servers from Cisco, Dell, HPE, NVIDIA Lenovo and Supermicro. Tell us what you need and we'll come back with options within 24 hours.

Configure your server now →

Description

The Supermicro AS-4125GS-TNRT is a high-density 4U rackmount GPU server engineered for AI, deep learning, and high-performance computing workloads. It supports dual AMD EPYC 9004/9005 Series processors with up to 192 cores, 24 DDR5 DIMM slots for up to 6TB of memory, and up to 8 double-width PCIe 5.0 GPUs including NVIDIA H100, H200, and L40S. With 24 hot-swap 2.5" drive bays, dual 10GbE networking, and four 2000W Titanium redundant power supplies, this server delivers exceptional scalability and reliability for enterprise AI infrastructure.

Features

Dual AMD EPYC 9004/9005 Series processor support with up to 192 cores/384 threads
- Up to 8× double-width PCIe 5.0 GPUs with direct-attach architecture
- 24× DDR5 DIMM slots supporting up to 6TB of 6000 MT/s ECC memory
- 9× PCIe 5.0 expansion slots for maximum I/O flexibility
- 24× hot-swap 2.5" drive bays (NVMe/SATA/SAS) plus M.2 slot
- 4× 2000W 80 PLUS Titanium redundant power supplies (2+2 or 3+1 configuration)
- 8× hot-swap heavy-duty PWM fans with intelligent thermal management
- Dual 10GbE RJ45 networking with AIOM/OCP 3.0 expansion support
- IPMI 2.0 with KVM-over-IP and dedicated management LAN
- TPM 2.0, Secure Boot, and NIST 800-193 compliant Silicon Root of Trust
- Redfish API support for modern datacenter automation
- Support for NVIDIA NVLink and AMD Infinity Fabric GPU interconnects
- Tool-less drive trays and rail kit included for easy deployment
- SuperDoctor 5 and Supermicro Update Manager (SUM) utilities

Warranty

All products sold by XS Network Tech include a 12-month warranty on both new and used items. Our in-house technical team thoroughly tests used hardware prior to sale to ensure enterprise-grade reliability.

All technical data should be verified on the manufacturer data sheets.

View full details

specs-tabs

Collapsible content

Technical Specifications

FAQs

Technical Specifications

Form Factor: 4U rackmount (17.2" W × 7" H × 29" D)
- Processors: Dual-socket AMD EPYC 9004/9005 Series, up to 400W TDP per CPU
- Memory: 24× DDR5 DIMM slots, 6000 MT/s, up to 6TB ECC RDIMM/LRDIMM
- GPU Support: Up to 8× double-width full-length PCIe 5.0 GPUs (NVIDIA H100, H200, L40S, A100, AMD Instinct MI210/MI300)
- Storage: 24× hot-swap 2.5" drive bays (NVMe/SATA/SAS), 1× M.2 slot
- Network: 2× 10GbE RJ45 LAN ports, 1× dedicated IPMI LAN, AIOM/OCP 3.0 support
- Expansion: 9× PCIe 5.0 slots optimized for GPU direct-attach configuration
- Power Supply: 4× 2000W redundant (2+2 or 3+1) 80 PLUS Titanium PSUs
- Cooling: 8× hot-swap heavy-duty PWM fans with optimal speed control
- Management: IPMI 2.0, KVM-over-IP, Redfish API, TPM 2.0, Secure Boot
- Operating Temperature: 10°C to 35°C

FAQs

Q: What GPU configurations are supported on the AS-4125GS-TNRT?
A: The AS-4125GS-TNRT supports up to 8 double-width full-length PCIe 5.0 GPUs including NVIDIA H100, H200, L40S, RTX 6000 Ada, A100, A40, and AMD Instinct MI210/MI300 series. Both active-cooled and passive-cooled GPUs are supported. The system uses direct-attached PCIe 5.0 x16 interconnects for maximum GPU bandwidth.

Q: How much memory can this server support?
A: The server supports up to 6TB of DDR5 ECC memory across 24 DIMM slots (12 per processor). It supports DDR5-6000 MT/s RDIMM and LRDIMM modules, providing high-bandwidth memory for data-intensive AI and HPC workloads.

Q: What storage configurations are available?
A: The AS-4125GS-TNRT features 24 hot-swap 2.5" drive bays supporting NVMe, SATA, and SAS drives. The front configuration includes 2 SATA bays and 4 dedicated NVMe bays, with additional expansion options. An onboard M.2 slot is also available for boot drives or caching.

Q: Is this server suitable for multi-GPU AI training workloads?
A: Yes, the AS-4125GS-TNRT is optimized for multi-GPU AI and deep learning training. With support for up to 8 high-performance GPUs, PCIe 5.0 direct-attach architecture, dual high-core-count AMD EPYC processors, and up to 6TB of fast DDR5 memory, it provides the compute density and bandwidth required for large-scale model training and inference.

Q: What networking options are included?
A: The server includes dual 10GbE RJ45 LAN ports for high-speed data connectivity and a dedicated IPMI LAN port for out-of-band management. AIOM (Add-In OCP Module) and OCP 3.0 expansion slots provide flexible networking upgrades including 25GbE, 100GbE, and InfiniBand options for high-performance cluster environments.

  • Request a Quote

    Looking for competitive pricing? Submit a request, and our team will provide a tailored quote that fits your needs.

  • Contact Us Directly

    Have a question or need immediate assistance? Call us for expert advice and real-time support.

    Call us Now  
  • Contact Us Directly

    Have a question or need immediate assistance? Call us for expert advice and real-time support.

    Contact us