Uvation MarketplaceMarketplace

Dell Power Edge XE9712 AI Server

A 1RU rack-scale AI compute sled pairing two NVIDIA Grace CPUs with four Blackwell Ultra GPUs, built for large-model training, inference and GB300 NVL72 deployments.

4x
GPUs per sled
1,152 GB
HBM3e
33 kW
Power shelf

Rack-Scale AI Built on
NVIDIA Grace Blackwell

Each 1RU sled pairs two Grace CPUs with four Blackwell Ultra GPUs, scaling to a full 72-GPU GB300 NVL72 rack.

Dell PowerEdge XE9712 front view

Scale from four Blackwell Ultra GPUs per sled to 72 GPUs and 36 Grace CPUs in a complete GB300 NVL72 rack.

Up to 900 GB/s of NVLink CPU-to-GPU bandwidth and 288 GB HBM3e per GPU support large-model training and inference.

A 33 kW rack power shelf with six 5500 W supplies feeds every sled over a shared 54 VDC busbar.

Dell PowerEdge XE9712 front view

Dimensions & Weight

Rack-level dimensions and weight for the Dell IR9048 48RU rack that hosts XE9712 compute sleds.

Dell PowerEdge XE9712 chassis
90.3 inches / 2,294 mm
Height
29.52 inches / 750 mm
Width
47.24 inches / 1,200 mm
Depth (52.76 in / 1,340 mm with stabilizers)
3,505.35 lbs / 1,590 kg (cabinet, wet)
Weight

Enterprise Management Built In

Maximize cluster availability with iDRAC10, advanced out-of-band monitoring, Redfish API automation, and comprehensive OpenManage tooling.

Embedded / At-the-Server Management

ASPEED AST2600 BMC · OpenBMC firmware · DC-SCI compatible architecture

Rack-Level Monitoring

Standards-based OpenBMC management · Rack-level monitoring · Management over the front RJ45 port

Security & Access Control

Silicon-based root of trust · Multi-factor authentication · Role-based access controls

Ports Overview

Network: 1 × 1GbE LOM; 1 × BlueField-3 SuperNIC Front: 1 × USB 3.0 · 1 × RJ45 management · 1 × Mini-DisplayPort · 1 × USB 2.0 Micro Type-B · 1 × 1GbE LOM Rear: N/A · Internal: N/A

Direct-to-Chip Liquid Cooling

Direct liquid cooling manages the Grace CPUs and Blackwell Ultra GPUs, while eight sled fans cool remaining components, sustaining performance in dense rack-scale deployments.

Dell PowerEdge XE9712 cooling and power
33 kW
Power shelf capacity
6x 5500W
AC power supplies

The rack, itemized

SUBSYSTEM

SPECIFICATION / CONFIGURATION DETAILS

Processor

Two NVIDIA Grace with 72 cores per processor

Operating System

Grace OS

Chipset

Not specified (NA) [1]

Accelerators / GPU

4 × NVIDIA Blackwell Ultra GPUs; 900 GB/s NVLink-C2C CPU-to-GPU interconnect

Memory

480 GB of LP DDR5 memory with ECC / CPU ; 288 GB HBM3e / GPU

Primary Storage

Up to 8 x EDSFF E1.S hot-swappable NVMe drives

Secondary Storage

1 × M.2 boot drive

Security

Silicon-based root of trust anchors end-to-end boot resilience while Multi-Factor Authentication (MFA) and role-based access controls safeguard trusted operations

Chassis

1RU compute sled installed in Dell IR9048 48RU rack

Power Supply

IR9048 rack consists of Power Shelf 33kW (PS33) that supports 6 x 5500 W AC PSUs

Ports

1 x USB 3.0 Type - A port, 1 x BIOS Serial out, 1 x RJ45 dedicated BMC port / management port, 1 x Mini-Display Port, 1 x 1Gb LOM, 4 x OSFP

Slots / PCIe Architecture

Up to 2 x16 FH HL Gen5, front risers only; 2 x Mezzanine - Gen5/Gen6, PCIe Cards : BF3 B3240 Super-NIC Full Height Half Length

Bezel

Not available (NA) [1]

Form Factor

1RU compute sled for IR9048 48RU rack

Network Card

1x 1GbE LOM

Warranty

Dell Limited Hardware Warranty; ProSupport / ProSupport Plus Options, Duration Subject to Order

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.