Supermicro AS-8125GS-TNMR2 — 8x MI300X Liquid-Cooled (EPYC 9004)
Direct-liquid-cooled 8-GPU MI300X density for sustained large-scale training.
We help you choose, configure, and deliver the right system — no obligation.
Configuration at a Glance
Tailored per engagement. Full technical overview below.
Configuration Options
Core specifications for this system. Every component is configurable to your workload — request a quote for a tailored build.
8x AMD Instinct MI300X OAM, 192GB HBM3 each (1.5TB total)
Dual AMD EPYC 9004 (Genoa)
Up to 6TB DDR5 (24 DIMM slots)
16x hot-swap NVMe + 2x SATA + 2x M.2 NVMe
Overview
The Supermicro AS-8125GS-TNMR2 packs eight MI300X OAM accelerators and dual AMD EPYC processors into an 8U chassis with direct-to-chip liquid cooling for uninterrupted, high-clock training. Nexus Compute configures the cooling loop, 1:1 GPU-to-NIC fabric, and ROCm stack, then validates thermals before warranty-backed delivery through authorized channels.
Who This Solution Is For
Business Benefits
Sustained training performance
Direct-to-chip liquid cooling holds peak clocks through long jobs that throttle air-cooled systems.
Higher rack density
Liquid cooling concentrates more accelerated compute per rack and improves overall facility efficiency.
1:1 GPU-to-NIC fabric
A dedicated NIC per GPU enables full-bandwidth RDMA scale-out across nodes.
Typical Business Use Cases
Multi-day foundation and large-model training
Dense liquid-cooled training rack build-outs
Distributed training over GPUDirect RDMA
High-throughput ROCm HPC workloads
Industry Applications
Technical Overview
Based on the Supermicro AS-8125GS-TNMR2 8U platform with eight MI300X OAM accelerators on a direct-to-chip liquid cooling loop and 4th-gen AMD Infinity Fabric for all-to-all GPU links. Dual AMD EPYC 9004 CPUs, up to 6TB of DDR5, and a 1:1 GPU:NIC topology drive distributed-training bandwidth.
| Accelerator | 8x AMD Instinct MI300X OAM, 192GB HBM3 each (1.5TB total) |
| GPU Interconnect | AMD Infinity Fabric, all-to-all (6.4TB/s bisection) |
| CPU | Dual AMD EPYC 9004 (Genoa) |
| System Memory | Up to 6TB DDR5 (24 DIMM slots) |
| Storage | 16x hot-swap NVMe + 2x SATA + 2x M.2 NVMe |
| Networking/Fabric | 8x NICs, 1:1 GPU:NIC for 400G RDMA scale-out |
| Cooling | Direct-to-chip (DLC) liquid cooling |
| Form Factor | 8U rackmount |
| Warranty | Nexus-managed OEM warranty (configurable term) |
Specifications are indicative and configured to each engagement. Request a quote for a configuration tailored to your requirements.
Warranty, Support & Fulfillment
Every system ships from an authorized channel, configured and tested, with the documentation enterprise buyers need — backed by warranty and a dedicated account team.
Enterprise Warranty
Full manufacturer warranty with optional on-site, next-business-day support and extended coverage.
Authorized Channel
Sourced through Tier-1 distribution and OEM partners — never grey market. Asset & warranty records included.
Lead Time & Deployment
48-hour quotes, then configured, burn-in tested, and delivered on a committed schedule.
Nationwide Fulfillment
Coordinated logistics, rack-and-stack, and delivery wherever your infrastructure lives.
Frequently Asked Questions
What facility support does liquid cooling require?
Direct-to-chip cooling needs facility coolant distribution (CDU and manifolds) and a rear-door or rack loop. We assess your data center and specify the CDU, flow, and redundancy as part of scoping.
Why liquid-cool MI300X instead of air?
At eight 750W accelerators per node, liquid cooling sustains higher clocks and rack density on long training runs. Air-cooled options like the Dell XE9680 suit facilities without liquid infrastructure.
Can multiple nodes form a training cluster?
Yes. The 1:1 GPU-to-NIC design supports 8-rail RDMA fabrics for multi-node scale-out. We design the fabric to your node count and growth plan.
Hardware Assistance
Configure the Supermicro AS-8125GS-TNMR2 — 8x MI300X Liquid-Cooled (EPYC 9004) with Nexus Compute
Tell us your requirements and a hardware specialist will help you specify, configure, and quote the right system — typically within two business days. No obligation.