SLYD
Read first

The 5-step pipeline.

How energy turns into a deployed, financed, offtake-matched cluster, and what SLYD does at every step.

How SLYD works →
Hardware

New and recovered GPU systems.

NVIDIA and AMD systems through documented manufacturer and qualified channel supply, with financing and deployment coordinated on the same platform.

Explore GPU hardware →
Marketplace

Compute, hardware, and power in one book.

Browse available GPU capacity by accelerator, configuration, region, and price, or bring supply to qualified demand.

Open marketplace →
Pre-qualify

Start with an indicative structure.

Tell us deal size, structure, and offtake. Any range is preliminary and subject to underwriting, diligence, and documentation.

Open Configure →
From the blog

GPU market trends and deployment playbooks.

Infrastructure best practices, hardware comparisons, and industry analysis from the SLYD team.

Read the blog →

Library

Manufacturer family

Dell PowerEdge AI servers

Dell's AI portfolio spans conventional 4U to 10U GPU servers, OCP rack-unit systems, and full rack-scale platforms. Which one fits is usually decided by the rack standard and cooling the site can take, not by the accelerator.

What are Dell PowerEdge AI systems?

Dell PowerEdge AI systems include GPU-accelerated servers and rack-scale platforms for training, inference, HPC, and enterprise AI workloads. The right configuration depends on the current PowerEdge model, accelerator, CPU architecture, memory, interconnect, storage, cooling, power, software, and service requirements.

Portfolio shape

Three shapes of Dell AI system

Dell's AI models fall into three groups that have genuinely different deployment consequences. Identifying the group first narrows the model list quickly.

Rack scale

Integrated NVL rack systems

XE9712 as a GB300 NVL72 system at 48U, shipping in an IR9000 rack. The whole rack is the product, and the facility has to accept it as delivered.

Dense 8-way

8-GPU accelerated servers

XE9780, XE9780L, XE9785, XE9680L, and XE9680. Conventional or OCP form factors carrying an 8-accelerator baseboard, deployed as nodes in a cluster.

Flexible

PCIe and mixed-use servers

XE7745 and similar 4U systems positioned by Dell for generative AI fine-tuning and inferencing, where a full 8-way baseboard is more than the workload needs.

Manufacturer catalog

Current Dell AI models by accelerator

Reproduced from Dell's PowerEdge AI servers catalog as published. Model lineups change; confirm the current configuration with Dell before relying on any row.

Source: Dell PowerEdge AI servers catalog, reviewed 18 August 2026. OU refers to the OCP rack unit standard, which requires a matching rack.
Model Form factor Processors Accelerators
PowerEdge XE9712 (GB300 NVL72) 48U, ships in IR9000 36 x Arm-based NVIDIA Grace 72 x NVIDIA B300
PowerEdge XE8712 (GB200 NVL4) 1OU, ships in IR7000 2 x Arm-based NVIDIA Grace 4 x NVIDIA B200
PowerEdge XE9780 10U 2 x Intel Xeon 6 6700P 8 x NVIDIA HGX B200 or B300
PowerEdge XE9780L 3OU 2 x Intel Xeon 6 6900P or 6700P 8 x NVIDIA HGX B300
PowerEdge XE9785 1OU 2 x AMD EPYC 9005 8 x AMD Instinct MI355X
PowerEdge XE9680L 4U 2 x 5th Generation Intel Xeon Scalable 8 x NVIDIA HGX B200
PowerEdge XE9680 6U 2 x 5th Generation Intel Xeon Scalable 8 x NVIDIA H100 or H200, 8 x AMD Instinct MI300X, or 8 x Intel Gaudi 3
PowerEdge XE7745 4U 2 x AMD EPYC 9005 series Positioned by Dell for generative AI fine-tuning and inferencing

Dell's catalog includes further AI models whose accelerator configuration could not be read unambiguously at review time. They are omitted here rather than described from inference. For anything not listed, ask for the current Dell specification sheet for the exact model.

Selection

Working from the site backwards

Dell publishes several systems with the same accelerator in different physical formats. That is the useful part of the portfolio, and it is why the facility question should be answered before the model is chosen.

  1. Which rack standard does the site use

    Several Dell AI models are published in OU form factors, which are OCP rack units and need a matching rack. A conventional 19-inch estate narrows the list immediately.

  2. Air or direct liquid

    Dense 8-way and rack-scale systems concentrate heat. Confirm the cooling design Dell publishes for the exact model, then confirm the facility can support it.

  3. CPU architecture

    Dell publishes Intel Xeon, AMD EPYC, and Arm-based NVIDIA Grace across the AI portfolio. This affects the software stack, not just the procurement line.

  4. Accelerator

    Once the physical constraints are settled, the accelerator choice is usually between two or three models rather than the whole catalog. See Blackwell, Hopper, and AMD Instinct.

Who provides what

Dell services and SLYD process are separate things

Dell

Manufacturer services

Dell publishes its own support, service, and management offerings. Their scope, term, regional availability, and eligibility are defined by Dell for the specific product and contract, and they should be read from Dell's current terms rather than from any reseller summary.

SLYD

Sourcing process

SLYD evaluates a stated requirement, identifies which documented sourcing paths are available for the requested system, and confirms configuration, seller authority, provenance, condition, availability, and commercial terms for the specific transaction.

SLYD is an authorized reseller for Dell Technologies, HPE, Supermicro, Lenovo, Cisco and IBM, and sources through authorized distribution including Arrow, Ingram Micro and TD SYNNEX. Configuration availability and lead times are quoted at time of order. SLYD does not claim factory-direct status, does not resell or extend Dell services, and does not present a manufacturer award or recognition as a SLYD credential. Counterparties and governing documents are disclosed to the buyer before a transaction proceeds.

What we need from you

To request a configuration

  • Workload and scale
  • Preferred accelerator, if already chosen
  • Node or rack count
  • Deployment location
  • Rack standard in use, 19-inch or OCP
  • Available rack power and cooling method
  • Target timeline
  • Condition tolerance, new or qualifying recovered
  • Compliance and export constraints
  • Capital structure preference
FAQ

Dell PowerEdge questions

Which current Dell systems support NVIDIA Blackwell accelerators?

Dell publishes PowerEdge XE9780 with 8 NVIDIA HGX B200 or B300, XE9780L with 8 NVIDIA HGX B300, XE9680L with 8 NVIDIA HGX B200, XE9712 as a GB300 NVL72 rack system with 72 NVIDIA B300 GPUs, and XE8712 as a GB200 NVL4 system with 4 NVIDIA B200 GPUs. Confirm the current configuration with Dell before relying on any of these details.

Which Dell systems support AMD Instinct accelerators?

Dell publishes PowerEdge XE9785 with 2 AMD EPYC 9005 processors and 8 AMD Instinct MI355X accelerators in a 1OU form factor, and publishes PowerEdge XE9680 with 8 AMD Instinct MI300X among its accelerator options.

What do the L suffix and OU form factors in Dell model names mean?

They signal a different physical system rather than a different accelerator. Dell publishes XE9780 at 10U and XE9780L at 3OU, and publishes XE9712 at 48U shipping in an IR9000 rack and XE8712 at 1OU shipping in an IR7000 rack. OU refers to the OCP rack unit standard, so the rack itself has to match. Confirm the rack standard and cooling design with Dell for the exact model quoted.

Does a Dell model listed on this page mean SLYD has it in stock?

No. Model names are reproduced from Dell's published catalog to explain the architecture choices available. They are not a statement of SLYD inventory, allocation, or availability. Whether a specific configuration can be obtained, in what quantity, and on what terms is confirmed at the transaction level.

Is SLYD a Dell partner or authorized reseller?

SLYD is an authorized reseller for Dell Technologies, HPE, Supermicro, Lenovo, Cisco and IBM, and sources through authorized distribution including Arrow, Ingram Micro and TD SYNNEX. Configuration availability and lead times are quoted at time of order. SLYD makes no claim of factory-direct status on this site. SLYD evaluates a stated buyer requirement and identifies which documented sourcing paths are available for the requested system, and the counterparties and governing documents for any transaction are disclosed to the buyer before it proceeds.

How do Dell service offerings relate to what SLYD provides?

They are separate. Dell publishes its own support and service offerings, and their scope, term, regional coverage, and eligibility are defined by Dell for the specific product and contract. SLYD does not resell, extend, or guarantee any Dell service, and any SLYD involvement is limited to the scope written into the executed documents for a specific project.

What information is needed to request a Dell configuration?

The workload and scale, preferred accelerator if already chosen, node or rack count, deployment location, available rack power and cooling method, target timeline, condition tolerance for new or recovered supply, and any facility, compliance, or capital constraints that shape the design.

Can Dell systems be evaluated for equipment financing?

SLYD originates and arranges eligible equipment financing and operating lease opportunities with third-party capital providers. Approval, pricing, collateral, term, and documentation depend on the borrower, equipment, deployment, repayment source, and transaction structure. Financing is not guaranteed.

Sources for the Dell information on this page

Model names, form factors, processors, and accelerator options are Dell's, reproduced as published on the review date. They are not SLYD guarantees of configuration, service, or availability. Dell is a trademark of its owner and is used here descriptively.

Request a Dell configuration

Share the workload, node or rack count, site, rack standard, and cooling method. That is enough to narrow the model list and start on sourcing paths.

Page updated: August 18, 2026

Reconnecting to the server...

Please wait while we restore your connection

An unhandled error has occurred. Reload 🗙