NVIDIA H100 NVL 94GB HBM2e PCIe Dual-GPU Module

NVIDIA H100 NVL 94GB HBM2e PCIe Dual-GPU Module

Brand: Dell | Category: GPUs

SKU: 900-21001-0060-000 | Part #: 900-21001-0060-000 | MPN: 900-21001-0060-000

Contact for Pricing — Request a Quote

Request a Quote Contact Us

About the NVIDIA H100 NVL 94GB HBM2e PCIe Dual-GPU Module

This dual-slot PCIe passive module delivers two NVIDIA H100 Hopper GPUs in a compact, datacenter-optimized form factor cooled entirely by passive airflow. Dell's NVIDIA H100 NVL 94GB HBM2e PCIe module (part number 900-21001-0060-000) is purpose-built for enterprise PCIe servers where space and thermal efficiency are critical, offering 94 GB of unified HBM2e memory across both processors and direct GPU-to-GPU connectivity via NVIDIA NVLink for seamless multi-GPU workloads.

The module connects via PCIe Gen5 x16 to the host system, providing 128 GB/s of bidirectional bandwidth for rapid data movement between server memory and GPU accelerators. Each H100 GPU features 4th Generation Tensor Cores supporting full-precision and mixed-precision compute across FP64, FP32, TF32, BF16, FP16, FP8, and INT8 formats, with Multi-Instance GPU (MIG) partitioning for flexible workload isolation. The 600 GB/s intra-module NVLink bandwidth (300 GB/s per direction) enables tightly coupled GPU clusters ideal for large language models, scientific simulation, and data-parallel inference. With a maximum thermal design power of 350 W per module and passive cooling architecture, this Dell accelerator eliminates active blower complexity while maintaining enterprise reliability. AI infrastructure teams and IT procurement departments seeking high-density GPU compute in existing PCIe server platforms will find this module a natural fit for production AI, HPC, and analytics environments. Contact Omnixon Global to request a quotation for the Dell NVIDIA H100 NVL series.

Key Specifications

  • Manufacturer: Dell
  • Manufacturer Part Number: 900-21001-0060-000
  • GPU Architecture: NVIDIA Hopper
  • Number of GPUs: 2 (Dual H100 PCIe GPUs)
  • Total Memory: 94 GB HBM2e
  • Memory per GPU: 47 GB HBM2e
  • GPU Interconnect: NVIDIA NVLink – 600 GB/s total (300 GB/s per direction)
  • Host Interface: PCIe Gen5 x16 – 128 GB/s bidirectional
  • Form Factor: Dual-slot PCIe passive module
  • Maximum TDP: 350 W per module

Technical Specifications

BrandDell
CategoryGPUs
SKU900-21001-0060-000
Part Number900-21001-0060-000
ConditionNew
Manufacturer Part Number900-21001-0060-000
GPU ArchitectureNVIDIA Hopper
Number of GPUs on Module2 (Dual H100 PCIe GPUs)
Total Memory94 GB HBM2e
Memory per GPU47 GB HBM2e
GPU InterconnectNVIDIA NVLink (between the two onboard GPUs)
Host InterfacePCIe Gen5 x16
Form FactorDual-slot PCIe passive module
CoolingPassive (datacenter airflow cooled)
Tensor Core Generation4th Generation (Hopper)
Supported PrecisionsFP64, FP32, TF32, BF16, FP16, FP8, INT8
Multi-Instance GPU (MIG)Supported
NVLink Bandwidth (intra-module)600 GB/s total (300 GB/s per direction)
PCIe Bandwidth128 GB/s bidirectional (PCIe Gen5 x16)
Max Thermal Design Power (TDP)350 W (per module)
Product LineNVIDIA H100 NVL Series
Target DeploymentEnterprise datacenter PCIe servers

Frequently Asked Questions about NVIDIA H100 NVL 94GB HBM2e PCIe Dual-GPU Module

What server platforms accept the NVIDIA H100 NVL 94GB HBM2e PCIe Dual-GPU Module?

Reference servers include Dell PowerEdge XE9680 / XE9712, HPE Cray XD670, Lenovo ThinkSystem SR685a / SR675 V3, Supermicro AS-A21GE / SYS-821GE, Gigabyte G593 / G894, ASUS ESC. Share your target platform in the RFQ and we will confirm chassis-to-GPU compatibility and recommended NIC pairing.

How long is the lead time on AI GPUs?

Highly model-dependent. L40S / RTX-class: typically 3-6 weeks. H100/H200/B200 in SXM form factor: 12-16 weeks for whole-platform allocations. We quote genuine-channel ETAs only — no grey-market promises.

Do you supply matched networking (Quantum InfiniBand / Spectrum-X)?

Yes — Omnixon stocks the full NVIDIA networking lineup (Quantum-2 / Quantum-X InfiniBand, Spectrum-X Ethernet, ConnectX NICs, BlueField DPUs) so we can quote a complete training-cluster BOM, not just the GPUs.

Can you help with NVIDIA AI Enterprise licensing?

Yes. We hold genuine-channels for NVIDIA AI Enterprise software subscriptions. Add it to your RFQ and we quote node-aligned licensing along with the hardware.