OpenMetal Hardware — Complete Index
Dedicated, single-tenant infrastructure at a fixed monthly rate — no hourly meter, no per-core cloud tax, included east-west bandwidth. Every hardware line, organized by generation and size.
Start Here
Find your hardware by need
Not sure of the product name? Pick by what you’re trying to do.
| Raw dedicated servers with full root access + IPMI | Bare Metal Dedicated → |
| A private cloud you fully control (OpenStack + Ceph, 3-node) | Hosted Private Cloud → |
| High-capacity storage (HDD + NVMe, Ceph capacity tier) | Storage Servers → |
| AI training, inference, or fine-tuning | GPU Servers → |
| Confidential computing / hardware-encrypted memory | TDX-enabled lines → |
| Compare OpenMetal against AWS on price/performance | Compare Hardware → |
| The engineering reasoning behind the hardware | Hardware Editorial → |
| A specific product question answered (specs, TDX, workloads, pricing) | OpenMetal Answers → |
Bare Metal Dedicated Servers
Single-tenant servers dedicated entirely to your workload — no shared tenancy, full root access and IPMI, boot/data drive isolation. Product line overview →
v5 — Intel Xeon 6 (Granite Rapids)
XL v5
High core count for dense virtualization, large in-memory workloads, and consolidation.
- 64 cores / 128 threads · Xeon 6530P
- 1 TB DDR5-6400 (to 2 TB)
- 4× 6.4 TB Micron 7500 MAX (25.6 TB)
- 20 Gbps private (to 40 Gbps)
Large v5
Balanced compute and memory for virtualization hosts, mid-to-large databases, mixed production.
- 32 cores / 64 threads · Xeon 6517P
- 512 GB DDR5-6400 (to 1 TB)
- 2× 6.4 TB Micron 7500 MAX (12.8 TB)
- 20 Gbps private (to 40 Gbps)
Medium v5
Entry server for containerized databases, mid-scale Kubernetes, and multi-threaded analytics.
- 24 cores / 48 threads · Xeon 6505P
- 256 GB DDR5-6400 (to 1 TB)
- 1× 6.4 TB Micron 7500 MAX NVMe
- 20 Gbps private (to 40 Gbps)
v4 — 5th Gen Intel Xeon Scalable
XXL v4
Maximum memory for very large in-memory databases and big-analytics workloads.
- 64 cores / 128 threads · Xeon Gold 6530
- 2 TB DDR5
- 20 Gbps private
XL v4
High-density compute. Xeon Gold 6530.
- 64 cores / 128 threads
- 1 TB DDR5
- 4× 6.4 TB NVMe · 20 Gbps private
XL v4 High-Frequency
Higher clocks for latency-sensitive and per-core-licensed workloads (trading, single-thread-bound DBs, game servers).
- 32 cores / 64 threads · Xeon Gold 6544Y
- 1 TB DDR5-5200
- 4× 6.4 TB NVMe (25.6 TB) · 20 Gbps private
Large v4
Balanced production server. Xeon Gold 6526Y.
- 32 cores / 64 threads
- 512 GB DDR5-5200
- 2× 6.4 TB NVMe · 20 Gbps private
Medium v4
Entry dedicated server. Xeon Silver 4510.
- 24 cores / 48 threads
- 256 GB DDR5-4400
- 1× 6.4 TB NVMe · 20 Gbps private
Hosted Private Cloud
Fully managed 3-node OpenStack + Ceph private clouds on dedicated hardware — your own cloud, root-level control, no per-instance metering. Product line overview →
v5 — Intel Xeon 6 (Granite Rapids)
XL v5
High-density private cloud for heavy VM density and large multi-tenant clouds.
- 3-node cluster · 192C / 384T
- 3 TB DDR5-6400
- 76.8 TB raw NVMe (~25.6 TB Ceph)
Large v5
Balanced private cloud for general production workloads.
- 3-node cluster · 96C / 192T
- 1.5 TB DDR5-6400
- 6× 6.4 TB Ceph NVMe · 60 Gbps mesh
Medium v5
Entry private cloud for dev/test clouds and small production OpenStack.
- 3-node cluster · 72C / 144T
- 768 GB DDR5-6400
- Ceph NVMe · 60 Gbps mesh (to 120)
v4 — 5th Gen Intel Xeon Scalable
XXL v4
Maximum-memory private cloud for memory-bound enterprise workloads.
- 3-node · Xeon Gold 6530
- 6 TB DDR5
- 115.2 TB NVMe cluster
Storage Servers
High-capacity storage nodes (bulk HDD + NVMe tier) for backup, archive, data lakes, and Ceph capacity tiers. Product line overview →
Storage Large v4
High-capacity storage node. Xeon Silver 4510.
- 24 cores / 48 threads · 256 GB DDR4
- 240 TB HDD + 25.6 TB NVMe
- 20 Gbps private
Storage Medium v4
Entry storage node. Xeon Silver 4510.
- 24 cores / 48 threads · 256 GB DDR5
- 120 TB HDD + 12.8 TB NVMe
- 20 Gbps private
GPU Servers
Dedicated NVIDIA GPU bare metal for AI training, inference, and fine-tuning — fixed monthly pricing, no per-GPU-hour meter. Contact OpenMetal for GPU pricing. Product line overview →
NVIDIA RTX PRO 6000 (Blackwell SE)
Inference, fine-tuning, and mixed AI/visual compute.
- 96 GB GDDR7 per GPU · 1–2 GPUs
- 1.6 TB/s memory bandwidth
- 2× Xeon 6530P · 64C / 128T · 1 TB DDR5
NVIDIA H200 NVL
Large-model training and inference, HPC.
- 141 GB HBM3e per GPU · 1–2 GPUs
- 4.8 TB/s memory bandwidth
- 2× Xeon 6530P · 64C / 128T · 1 TB DDR5
GPU Clusters
Dedicated multi-GPU infrastructure for AI training and inference at scale.
- Multi-node GPU clusters
- Contact for configuration & pricing
Add GPU to Existing Cloud
Attach GPU nodes to a running OpenMetal private cloud or bare metal deployment.
- Expansion path for existing customers
Cross-cutting
Compare Hardware
High-intent comparison pages, grouped by type.
OpenMetal vs AWS
OpenMetal vs OpenMetal
GPU Comparisons
Go Deeper
Hardware EditorialThe engineering and architecture reasoning behind the hardware — for platform architects and SREs.
Cross-cutting
OpenMetal AnswersProduct-specific Q&A entries. Grouped by hardware line, then by topic for cross-cutting questions.
By hardware line
Bare Metal and Hosted Private Clouds
v5 25
- Intel TDX vs Intel SGX on the OpenMetal XL v5
- Migrating Azure Confidential VM Workloads to OpenMetal XL v5
- How Many Cores Does the OpenMetal XL v5 Have
- Ceph Usable Storage on a 3-Node Hosted Private Cloud XL v5
- Intel TDX on the OpenMetal XL v5: Active Out of the Box
- Memory Configuration on the OpenMetal XL v5
- NVMe Drives in the OpenMetal XL v5 Server
- OpenMetal XL v5 vs AWS m7i.metal-48xl
- Intel Xeon 6517P on the OpenMetal Large v5
- OpenMetal Large v5 Memory Bandwidth
- OpenMetal Large v5 Drive Bay Capacity
- OpenMetal Large v5 NVMe Drive Specs
- OpenMetal Large v5 L3 Cache for Database and Analytics
- OpenMetal Large v5 Intel TDX Confidential Computing
- Upgrading a Deployed OpenMetal Large v5 to TDX
- Combining SGX and TDX on the OpenMetal Large v5
- HIPAA BAA for OpenMetal Large v5 TDX Workloads
- OpenMetal Large v5 HPC Ceph Usable Storage
- TDX VMs on OpenMetal Large v5 Hosted Private Cloud
- OpenMetal Large v5 vs AWS i4i Comparison
- OpenMetal Medium v5 Bare Metal Server Specifications
- NVMe Drives in the OpenMetal Medium v5 Bare Metal Server
- Maximum RAM in the OpenMetal Medium v5 Bare Metal Server
- Intel TDX Support on the OpenMetal Medium v5
- RAM Upgrade Required to Enable TDX on the OpenMetal Medium v5
v4 43
- OpenMetal XL v4 vs XXL v4: Key Differences
- NVMe Drives in the OpenMetal XXL v4 Server
- RAM Upgrades on a Deployed OpenMetal XXL v4
- Running Kubernetes on the OpenMetal XXL v4
- Best Workloads for the OpenMetal XXL v4 Server
- In-Memory Database Performance on the OpenMetal XXL v4
- Maximum RAM Configuration on the OpenMetal XXL v4
- Ceph Storage Capacity in a 3-Node XXL v4 OpenMetal Cluster
- OpenMetal XXL v4 vs AWS x2idn: Dedicated vs Cloud
- OpenMetal XL v4 Bare Metal Server Pricing
- OpenMetal XL v3 vs XL v4: What Changed
- NVMe Drives in the OpenMetal XL v4 Server
- Best Workloads for the OpenMetal XL v4 Bare Metal Server
- Intel TDX on the OpenMetal XL v4: Enabled by Default
- OpenMetal XL v4 vs AWS i4i: Dedicated Bare Metal vs Cloud Instances
- DDR5-5200 vs DDR5-4800 on OpenMetal XL v4 Servers
- OpenMetal XL v4 vs XL v5: What Changed
- XL v4 vs XL v4 High Frequency: Which Server to Choose
- Is the XL v4 High Frequency Suitable for HFT Infrastructure?
- OpenMetal Large v3 vs Large v4 Comparison
- OpenMetal Large v4 NVMe Drive Specs
- Intel TDX on OpenMetal Large v4
- OpenMetal Storage Large v4 Capacity
- OpenMetal Large v4 vs AWS i4i Comparison
- OpenMetal Medium v4 vs Large v4 Comparison
- Storage Medium v4 vs Storage Large v4
- Medium v4 vs Large v4 Hosted Private Cloud
- Expanding Storage Medium v4 to Large v4
- OpenMetal Medium v4 TDX vs. Large v4 TDX: Which Should You Choose?
- OpenMetal Large v4 vs Large v5 Comparison
- OpenMetal Medium v4 Bare Metal Server Specs
- NVMe Drives in the OpenMetal Medium v4
- OpenMetal Medium v4 vs AWS m7i
- OpenMetal Storage Medium v4 Capacity
- OpenMetal Medium v4 Intel Xeon Processor
- Does the OpenMetal Medium v4 Support Intel TDX?
- What Is Required to Enable Intel TDX on the OpenMetal Medium v4?
- OpenMetal Medium v4 vs. Medium v4 TDX Edition: What’s the Difference?
- How Much RAM Does the OpenMetal Medium v4 Need for TDX?
- What Workloads Run Well on the OpenMetal Medium v4 TDX Edition?
- How to Upgrade an OpenMetal Medium v4 to Enable TDX
- OpenMetal Medium v5 vs Medium v4: Key Changes
GPU
H200 11
- Running a 70B LLM on a Single OpenMetal H200
- Building a Multi-GPU Cluster with OpenMetal H200s
- NVMe Storage in the OpenMetal H200 GPU Server
- The CPU Paired with the OpenMetal H200
- NVIDIA H200 vs H100: Key Differences
- Is the NVIDIA H200 Faster Than the H100 for Inference?
- Why OpenMetal Offers the H200 Instead of the H100
RTX PRO 6000 (RP6000) 11
- GPU Memory on the OpenMetal RP6000
- Training and Fine-Tuning on the OpenMetal RP6000
- Mixed RP6000 and H200 GPU Clusters on OpenMetal
- Attaching RP6000 GPU Nodes to an Existing Deployment
- NVIDIA RTX Pro 6000 vs L40S: Key Differences
- Is the RTX Pro 6000 Better Than the L40S?
- NVIDIA RTX Pro 6000 vs H100: Key Differences
- Choosing Between the OpenMetal RP6000 and H200
By topic (cross-cutting)
Migration 2
Networking 11
- How OpenMetal’s Hosted Private Cloud Networking Is Structured
- OpenMetal’s Networking Model: Software-Defined, Hardware VLAN, or Both
- How OpenMetal Separates Infrastructure, Storage, and Tenant Network Traffic
- OpenMetal Dedicated Networking vs. Shared Public Cloud Bandwidth
- Internal Node-to-Node Network Throughput in an OpenMetal Private Cloud
- OpenMetal Network Port Speeds for Bare Metal and Private Cloud
- Private Bandwidth Included with OpenMetal Bare Metal Servers
- Dedicated VLANs for OpenMetal Bare Metal Servers
- VxLAN Support on OpenMetal Bare Metal Servers
- Sharing VLANs Between Bare Metal and Private Cloud
- OpenMetal Egress Pricing vs AWS Per-GB
Other Hardware Answers 16
- Kubernetes on OpenMetal Bare Metal Servers
- Boot and Data Drive Isolation on OpenMetal
- RAM Upgrades on Deployed OpenMetal Servers
- OpenMetal Proxmox Bare Metal Reference Architecture
- OpenMetal Storage Servers for Ceph Clusters
- OpenMetal Storage and Compute Node Connectivity
- OpenMetal vs AWS Dedicated Server Cost
- OpenMetal Storage Pricing vs AWS S3 and EBS
- Intel Speed Select Technology on OpenMetal Bare Metal Servers
- When CPU Clock Speed Beats Core Count on Bare Metal
- OpenMetal vs AWS for 2TB RAM Server Cost
- Intel Xeon Silver 4510 vs Xeon 6505P: Architecture Comparison
- OpenMetal vs AWS Cost for Dedicated Granite Rapids
- OpenMetal Persistent NVMe vs AWS i4i Ephemeral Storage
- What FP4 (NVFP4) Is and Why It Matters
- GDDR7 vs HBM3 for AI Training and Inference
Platform & OpenStack 3
Pricing & Billing 4
Processors, Memory & Specs 2
Security & Confidential Computing 9
- Intel TDX vs SGX on OpenMetal Servers
- OpenMetal Servers with TDX Enabled by Default
- OpenMetal HIPAA Compliance for Healthcare
- Intel TDX vs SGX on OpenMetal Bare Metal Servers
- Running SGX Enclaves and TDX VMs on the Same OpenMetal Server
- RAM Requirement for Intel TDX on OpenMetal Bare Metal
- How Intel TDX Remote Attestation Works on OpenMetal Bare Metal
- OpenMetal BAA Availability for HIPAA Workloads
Not sure which line fits?
Talk to an OpenMetal engineer about your workload, or spin up a proof of concept to test on real hardware.