Uvation MarketplaceMarketplace

HPE ProLiant Compute XD685 DLC High Power Server

The HPE ProLiant Compute XD685 DLC High Power Server combines 8 NVIDIA B300 SXM GPUs, dual 5th Generation AMD EPYC processors, high-capacity DDR5 memory, and direct liquid cooling in a 5U rack server.

8×
NVIDIA B300 SXM GPUs
5U
direct-liquid-cooled chassis
Up to 6.1 TB
DDR5 memory

High-Density AI Compute with Direct Liquid Cooling

Designed for high-performance AI computing, the XD685 combines NVIDIA B300 GPUs, dual AMD EPYC processors, and direct liquid cooling in a rack-mounted platform.

NVIDIA B300 GPU Acceleration

Features 8 NVIDIA B300 SXM GPUs, each with 288 GB of memory.

High-Capacity DDR5 Memory

Supports 24 DDR5 DIMMs with up to 6.0 TB or 6.1 TB of memory using 256 GB or 128 GB RDIMMs, at speeds up to 6400 MT/s.

Redundant Titanium Plus Power

Equipped with six 3000W power supplies operating at 54V in an N+1 configuration, with 200–277 VAC input.

NVIDIA Blackwell Ultra
GPU Architecture

Built for GPU-intensive computing, the XD685 combines eight NVIDIA B300 SXM accelerators with dual 5th Generation AMD EPYC processors and high-capacity DDR5 memory in a direct-liquid-cooled platform.

HPE ProLiant Compute XD685 DLC chassis, front view
8× NVIDIA B300 SXM
GPUs
288 GB HBM3e
GPU Memory
Dual 5th Generation AMD EPYC
CPUs
Up to 6.1 TB DDR5 Memory
Memory

A platform for models that refuse to fit the old shape.

Built for compute-intensive AI pipelines, the HPE ProLiant Compute XD685 DLC supports large model development, specialized model adaptation, and GPU-accelerated deployment in a liquid-cooled platform.

Foundation Model Training

Run large-scale training workloads across eight NVIDIA B300 GPUs, with high-bandwidth GPU memory for demanding compute tasks.

Domain-Specific Model Tuning

Adapt pretrained models for enterprise datasets, specialized applications, and industry-specific AI use cases.

High-Throughput AI Inference

Deploy GPU-accelerated inference pipelines for generative AI, language models, and deep learning applications.

Engineered to dominate

GPU Density Meets Liquid-
Cooled Design

The XD685 combines eight NVIDIA B300 GPUs, dual AMD EPYC processors, and direct liquid cooling in a 5U chassis, providing a foundation for demanding AI compute environments.

Exploded view of the HPE ProLiant Compute XD685 liquid-cooled chassis
8× NVIDIA B300
GPUs
288 GB HBM3e
GPU Memory
Dual 5th Generation AMD EPYC processor
CPUs
5U direct-liquid-cooled chassis
Form Factor
Exploded view of the HPE ProLiant Compute XD685 liquid-cooled chassis
8× NVIDIA B300
GPUs
288 GB HBM3e
GPU Memory
Dual 5th Generation AMD EPYC processor
CPUs
5U direct-liquid-cooled chassis
Form Factor

The rack, itemized.

SUBSYSTEM

SPECIFICATION / CONFIGURATION DETAILS

Processor

2x 5th Generation AMD EPYC Processors

Operating System

Canonical Ubuntu Server LTS, Red Hat Enterprise Linux (RHEL), and SUSE Linux Enterprise Server (SLES)

Chipset

HPE iLO 6 ASIC

Accelerators / GPU

8× NVIDIA B300 SXM; 288 GB per GPU

Memory

24 DDR5 DIMMs Up to 6.0 TB / 6.1 TB using 256 GB or 128 GB RDIMMs running at speeds up to 6400 MT/s

Primary Storage

2 x M.2 NVMe SSDs

Secondary Storage

8 × E3.S NVMe SSDs

Security

HPE iLO 6 Silicon Root of Trust; Secure Boot and Secure Erase; Cryptographically signed firmware validation; Embedded Trusted Platform Module (TPM 2.0) support

Chassis

5U direct-liquid-cooled chassis

Power Supply

6x 3000W @54V Titanium Plus (N+1), 200-277 VAC input

Ports

iLO Ethernet; VGA; USB; configurable network ports

Slots / PCIe Architecture

8 × HHHL PCIe Gen5 x16; 3 × FHHL PCIe Gen5 x16

Bezel

N/A

Form Factor

5U rack server

Network Card

Configurable Ethernet and InfiniBand adapters

Warranty

3/3/3 warranty - three years parts, three years labor, and three years onsite support

Front view of the HPE ProLiant Compute XD685

Power Fabric

6× 3000W power supplies

54V Titanium Plus

Nuclear-Ready, Plug and Play AI Factories

Bypass grid constraints with self-contained, high-density inference ready infrastructure directly adjacent to clean, base load nuclear power.

Global Deployment Engine

Any Infra, Anywhere in the world.

Global Shipping and Deployment Capability treats advanced infrastructure as mobile, modular freight.

Packaged complex computing nodes, cooling manifolds, and power distribution units into certified, standard form factors.

360° Service Ecosystem

Unifying Legacy with Next-Gen Compute.

Vertically integrated service ecosystem that serves as an operational bridge between legacy environments and next-generation, AI-native infrastructure.

Finance and Leasing

Capital-as-a-Service for High-Density Compute

Convert upfront capital expenditures (CapEx) into predictable operating expenses (OpEx).

Uvation allows enterprises and sovereign entities to deploy multi-megawatt AI factories and advanced GPU clusters without liquidity constraint.