Skip to content
HomeSolutionsGPU ServersLenovo ThinkSystem SR685a V3 — 8x H200 SXM, AMD EPYC 9005
Nexus Compute

Lenovo ThinkSystem SR685a V3 — 8x H200 SXM, AMD EPYC 9005

AMD EPYC Turin paired with 8x H200 for high-throughput generative AI.

Full manufacturer warrantyAuthorized channel48-hour quote

We help you choose, configure, and deliver the right system — no obligation.

Lenovo ThinkSystem SR685a V3 — 8x H200 SXM, AMD EPYC 9005 — Nexus Compute enterprise hardware
Lenovo ThinkSystem SR685a V3 — 8x H200 SXM, AMD EPYC 9005 hardware detail 1
Lenovo ThinkSystem SR685a V3 — 8x H200 SXM, AMD EPYC 9005 hardware detail 2
Lenovo ThinkSystem SR685a V3 — 8x H200 SXM, AMD EPYC 9005 hardware detail 3

Configuration at a Glance

GPU / Accelerator8× NVIDIA H200 SXM5, 141GB HBM3e each
GPU InterconnectNVLink + NVSwitch fully connected
CPU2× 5th Gen AMD EPYC 9005 'Turin', up to 128 cores each
MemoryTruDDR5 up to 6000MHz, 24 DIMM slots

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.

GPU / Accelerator

8× NVIDIA H200 SXM5, 141GB HBM3e each

Processor

2× 5th Gen AMD EPYC 9005 'Turin', up to 128 cores each

Memory

TruDDR5 up to 6000MHz, 24 DIMM slots

Storage

Up to 16× hot-swap PCIe Gen5 NVMe

Overview

The ThinkSystem SR685a V3 combines eight NVSwitch-linked NVIDIA H200 SXM GPUs with dual 5th Gen AMD EPYC 9005 'Turin' processors in an 8U chassis tuned for generative AI throughput. Nexus Compute specifies, configures, and tests each system to your memory, storage, and fabric targets, delivered warranty-backed through authorized Lenovo channels.

Who This Solution Is For

Teams wanting maximum CPU core count alongside 8x H200
Generative AI platforms serving and fine-tuning at scale
Organizations standardizing on AMD EPYC compute
Enterprises needing high PCIe Gen5 lane availability

Business Benefits

Massive CPU headroom

Dual EPYC 9005 Turin CPUs reach up to 256 cores so data pipelines never starve the eight H200 GPUs.

Generative AI throughput

NVSwitch-coupled H200 GPUs with fast PCIe Gen5 paths sustain high-volume training and inference.

Lenovo serviceability

XClarity management and Lenovo support bring enterprise lifecycle operations to dense GPU nodes.

Typical Business Use Cases

1

Generative and large language model fine-tuning

2

High-throughput production inference serving

3

CPU-intensive data preprocessing alongside GPU training

4

Mixed training and serving on a single node

Industry Applications

AI & Machine LearningSaaS & SoftwareFinancial ServicesMedia & EntertainmentHigher Education & Research

Technical Overview

An 8U platform on the NVIDIA HGX H200 baseboard with NVLink and NVSwitch, driven by dual 5th Gen AMD EPYC 9005 'Turin' CPUs with TruDDR5 memory at up to 6000MHz across 24 DIMM slots. PCIe Gen5 connectivity links GPUs, CPUs, NVMe, and networking, with up to 16 hot-swap NVMe drives.

GPU / Accelerator8× NVIDIA H200 SXM5, 141GB HBM3e each
GPU InterconnectNVLink + NVSwitch fully connected
CPU2× 5th Gen AMD EPYC 9005 'Turin', up to 128 cores each
MemoryTruDDR5 up to 6000MHz, 24 DIMM slots
StorageUp to 16× hot-swap PCIe Gen5 NVMe
Networking / Fabric400GbE or NDR InfiniBand adapters (configurable)
Form Factor8U rackmount, air-cooled
ManagementLenovo XClarity Controller
WarrantyLenovo support, 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

Why choose EPYC 9005 over a Xeon-based H200 node?

EPYC 9005 Turin offers very high core counts and PCIe Gen5 lane density, which benefits data-heavy preprocessing and serving pipelines feeding the GPUs. We help you weigh it against Xeon-based platforms based on your workload mix.

Can this node both train and serve?

Yes. The 1.1TB of pooled HBM3e and large CPU core count make it well suited to mixed fine-tuning and high-throughput inference on one node. We configure storage and networking for your blend.

What are the power and cooling requirements?

This is a high-power 8U air-cooled system for data center deployment. We confirm rack power, airflow, and thermal headroom during the quote.

Hardware Assistance

Configure the Lenovo ThinkSystem SR685a V3 — 8x H200 SXM, AMD EPYC 9005 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.