This page compares the OpenMetal Bare Metal Dedicated Server XXL v4 with the AWS x2idn.32xlarge and x2idn.metal — the closest AWS equivalents by RAM profile for high-memory, NVMe-accelerated workloads. Both the XXL v4 and the x2idn family offer 2048GB of RAM, but differ sharply on storage volume, storage persistence, tenancy model, confidential computing, and pricing structure. The XXL v4 ships with 38.4TB of persistent NVMe — five times the ephemeral NVMe available on x2idn instances — alongside Intel TDX active by default, single-tenant dedicated hardware, and a fixed monthly pricing model that does not meter by hour, IOPS, or egress byte.
For full XXL v4 hardware specifications, see the primary XXL v4 spec page. For egress cost modeling, see the 95th-percentile billing section below.
Key Takeaways
- 38.4TB persistent NVMe vs 7.6TB ephemeral storage — the XXL v4 ships with six Micron 7500 MAX 6.4TB NVMe drives; data survives reboots, OS reinstalls, and maintenance windows; x2idn instance store is ephemeral and lost on instance stop or terminate, requiring EBS for persistence at additional cost
- Dedicated physical tenancy vs shared vCPU allocation — on OpenMetal, 64 physical cores belong to one customer; x2idn.32xlarge runs 128 vCPUs under AWS’s Nitro hypervisor; x2idn.metal provides more physical access at a significant premium
- Intel TDX active by default vs no native TDX on x2idn — the XXL v4 ships TDX enabled at the base 2048GB configuration with no upgrade required; AWS x2idn does not offer Intel TDX; Nitro Enclaves are available for application-level isolation with a different attestation model
- Fixed monthly vs on-demand metered — OpenMetal billing is a flat monthly rate regardless of CPU utilization, storage IOPS, or runtime hours; AWS on-demand charges accumulate by the hour for the life of the instance
- 95th-percentile egress vs per-GB transfer — OpenMetal’s egress model does not penalize burst traffic; AWS charges $0.09/GB
- IPMI out-of-band management vs SSH/SSM only — OpenMetal provides full IPMI console access for OS reinstall and BIOS-level configuration; AWS instances are accessible only through the running OS or AWS Systems Manager
Spec Comparison
| Component | OpenMetal XXL v4 | AWS x2idn.32xlarge | AWS x2idn.metal |
|---|---|---|---|
| CPU | 2x Xeon Gold 6530 (Emerald Rapids) | Intel | Intel |
| Physical Cores | 64 dedicated | 128 vCPUs | 128 |
| RAM | 2048GB DDR5 4800MHz | 2048 GB | 2048 GB |
| NVMe Storage | 38.4TB persistent (6x Micron 7500 MAX 6.4TB) | 7.6TB ephemeral | 7.6TB ephemeral |
| Storage Persistence | Persistent across reboots | Ephemeral — lost on stop/terminate | Ephemeral — lost on stop/terminate |
| Network | 20Gbps private + 6Gbps public | Up to 100 Gbps | Up to 100 Gbps |
| Tenancy | Single-tenant, dedicated physical | Shared Nitro hypervisor | Bare metal Nitro |
| Remote Management | IPMI out-of-band console | SSH / AWS SSM only | SSH / AWS SSM only |
| Intel TDX | Active by default | Not available | Not available |
| Intel SGX | Yes — 128GB max EPC | N/A | N/A |
| Egress Model | 95th-percentile (burstable) | Per-GB ($0.09/GB) | Per-GB ($0.09/GB) |
| Pricing Model | Fixed monthly | On-demand hourly / Reserved | On-demand hourly / Reserved |
AWS instance selection: x2idn.32xlarge is the standard high-memory NVMe instance at 2048GB RAM; x2idn.metal provides direct hardware access on the same platform. The x2i family is AWS’s high-memory NVMe-optimized line.
Ready to Compare Costs for Your Workload?
Tell us about your workload — current AWS spend, egress volume, storage requirements — and we’ll provide a direct cost comparison against your actual usage.
Processor: Dedicated Cores vs Shared vCPUs
The OpenMetal XXL v4 provides 64 physical cores from two Intel Xeon Gold 6530 processors (Emerald Rapids, 32 cores per socket). Every core is allocated to one customer — no other tenant shares CPU time, cache, or memory bandwidth with the XXL v4’s workloads. BIOS configuration, CPU frequency governors, Intel Speed Select Technology profiles, and processor affinity settings are fully accessible via IPMI.
AWS x2idn.32xlarge instances present 128 vCPUs under the Nitro hypervisor. For workloads that depend on consistent per-core throughput — in-memory databases, large-scale OLTP, real-time analytics — physical core dedication removes a scheduling variable that vCPU allocation introduces. The x2idn.metal variant provides more direct hardware access but at a higher per-hour cost and without the fixed-cost model of the OpenMetal XXL v4.
The XXL v4’s Gold 6530 includes Intel AMX for hardware-accelerated matrix operations, AVX-512 for SIMD-heavy workloads, and Intel QAT for cryptographic acceleration — all available without hypervisor overhead on bare metal. Some CPU instruction extensions may be unavailable or throttled by hypervisor layers on AWS instances.
Memory: 2TB DDR5 for High-Memory Workloads
Both the XXL v4 and the x2idn family offer 2048GB of RAM. The difference is upgradeability, bandwidth, and confidential computing integration. On the OpenMetal XXL v4, RAM is physical DDR5 4800MHz ECC across 32 DIMM slots, delivering approximately 614 GB/s of aggregate memory bandwidth. Custom configurations above 2048GB (up to 8192GB) are available by contacting OpenMetal — an in-place upgrade that does not require stopping and restarting on different hardware.
On AWS, increasing memory beyond an instance’s fixed allocation requires migrating to a different instance type — a disruptive operation that involves stopping the instance and provisioning new hardware. There is no in-place RAM upgrade path on AWS instances.
The XXL v4’s 2048GB RAM configuration also activates Intel TDX by default, enabling hardware-isolated, memory-encrypted Trust Domains for confidential workloads. AWS x2idn does not support Intel TDX; Nitro Enclaves provide a different application-level isolation model with a different attestation chain.
Storage: 38.4TB Persistent NVMe vs 7.6TB Ephemeral
This is the most significant structural difference between the XXL v4 and the x2idn family.
Boot and data isolation
OpenMetal XXL v4 storage: Six Micron 7500 MAX 6400GB NVMe drives (38.4TB total) persist across server reboots, OS reinstalls, and maintenance events. Boot and data drives are on separate paths — an OS reinstall via IPMI does not touch the data NVMe pool. At full array throughput, the six drives deliver approximately 42,000 MB/s sequential read and 6.6 million random read IOPS.
AWS x2idn instance storage: The x2idn family includes 7.6TB of NVMe SSD instance store — approximately one-fifth of the XXL v4’s NVMe capacity — and the storage is ephemeral. Data is lost when the instance is stopped, terminated, or moved to a different host during AWS maintenance. For any workload requiring data persistence, x2idn customers must provision EBS volumes separately: adding per-GB-month and per-IOPS costs, adding network latency (EBS is network-attached storage), and adding a point of failure not present in the XXL v4’s direct-attached NVMe architecture.
Micron 7500 MAX NVMe data drives
NVMe performance comparison per drive (source: Micron 7500 NVMe SSD Tech Prod Spec Rev. A 10/2023):
| Metric | Per Drive | 6-Drive Array | AWS x2idn NVMe |
|---|---|---|---|
| Sequential Read | 7,000 MB/s | ~42,000 MB/s (6-drive est.) | N/A |
| Sequential Write | 5,900 MB/s | ~35,400 MB/s (6-drive est.) | N/A |
| Random Read IOPS | 1,100,000 | ~6,600,000 (6-drive est.) | N/A |
| Random Write IOPS | 400,000 | ~2,400,000 (6-drive est.) | N/A |
| QoS (99.9999%) | Sub-1ms | — | N/A |
| Persistence | Always-on | Always-on | Ephemeral (lost on stop) |
EBS cost for persistent storage to match the XXL v4: gp3 at $0.08/GB/month adds approximately $3,072/month for 38.4TB — a cost with no analog on the OpenMetal XXL v4 where all NVMe is included in the fixed monthly price.
Networking
The OpenMetal XXL v4 provides 20Gbps private bandwidth (LACP-bonded, unmetered) and 6Gbps public bandwidth billed on the 95th-percentile model. All east-west traffic between OpenMetal servers on the same VLAN is unmetered — Ceph replication, database synchronization, Kubernetes pod-to-pod traffic, and inter-service calls cost nothing regardless of volume.
AWS x2idn instances provide up to 100 Gbps of network bandwidth, but cross-AZ traffic is billed at $0.01/GB, and public egress is billed at $0.09/GB for the first 10 TB/month in us-east-1. For high-memory workloads that generate large egress — in-memory analytics platforms exporting results, database read replicas serving external clients, or distributed caches with broad query patterns — per-GB egress charges accumulate regardless of whether the traffic is bursty or sustained.
Egress pricing: 95th-percentile billing, not per-GB transfer.
OpenMetal measures egress in 5-minute intervals and discards the top 5%. The billing rate is calculated against the 95th-percentile sustained rate, not total bytes transferred. AWS us-east-1: $0.09/GB for the first 10 TB/month. A workload producing 50 TB/month of public egress costs approximately $4,500/month in AWS transfer charges alone. For blockchain nodes, media delivery, data pipeline APIs, and any workload with sustained or unpredictable egress, the per-GB model scales costs non-linearly in ways that 95th-percentile billing does not.
Security and Confidential Computing
OpenMetal XXL v4 ships Intel TDX enabled at the base 2048GB configuration — hardware-isolated Trust Domains with full-memory encryption are available immediately on the physical server, with no upgrade required. TDX on bare metal means the isolation boundary extends to the physical CPU: no hypervisor layer, no cloud provider firmware, no shared hardware management controller sits between the Trust Domain and the processor. Intel SGX with 128GB EPC is also available alongside TDX on the same server.
AWS x2idn does not support Intel TDX. Nitro Enclaves are available for application-level confidential computing on supported instance types, but operate within the Nitro hypervisor stack — the isolation model and attestation chain differ from Intel TDX on bare metal.
| Feature | OpenMetal XXL v4 | AWS x2idn.32xlarge | AWS x2idn.metal |
|---|---|---|---|
| Intel TDX | Enabled by default | Not available | Not available |
| Intel SGX | Yes — 128GB max EPC | N/A | N/A |
| TME-MK | Enabled | Not applicable | Not applicable |
| Nitro Enclaves | — | Available | Available |
| Physical isolation | Single-tenant hardware | Shared Nitro host | Nitro metal |
| IPMI access | Yes — full OOB | No | No |
HIPAA and Regulatory Compliance
OpenMetal holds a HIPAA BAA at the organizational level — one BAA covers all workloads running on OpenMetal infrastructure. The Ashburn deployment location is further strengthened by an NTT DATA facility that holds HIPAA as a facility-operator certification alongside SOC 1/2 Type II, ISO 27001, PCI DSS, and NIST 800-53 HIGH. Los Angeles, Amsterdam, and Singapore locations are HIPAA-eligible at the OpenMetal organizational level only. AWS also offers HIPAA BAAs, but HIPAA eligibility is per-service — not all AWS services are covered and customers must verify each service against the AWS HIPAA-eligible services list.
OpenMetal’s infrastructure-level HIPAA posture — BAA, dedicated physical tenancy, TDX isolation, and Ashburn facility-level HIPAA — provides a simpler compliance path for workloads running exclusively on XXL v4 hardware without consuming additional cloud-managed services.
When Each Platform Wins
When OpenMetal XXL v4 Wins
- High-memory workloads requiring large persistent NVMe — in-memory databases, SAP HANA, large Redis clusters, Oracle RAC; 38.4TB persistent NVMe vs 7.6TB ephemeral is a structural difference
- Sustained 24/7 production workloads — fixed monthly pricing is predictable; on-demand AWS pricing accumulates continuously
- Confidential computing at the hardware level — TDX active by default, SGX 128GB EPC, single-tenant hardware; x2idn has no native TDX
- Egress-heavy workloads — analytics exports, blockchain nodes, media APIs; per-GB charges at AWS scale quickly exceed OpenMetal’s 95th-percentile model
- Compliance and isolation requirements — single-tenant hardware, HIPAA BAA at org level, facility-level HIPAA in Ashburn
- Predictable budget planning — fixed monthly cost, lockable for up to five years
When AWS x2idn Is the Better Fit
- Scale-to-zero or short-lived workloads — batch jobs, dev/test environments, and event-driven processing that run for hours rather than months
- AWS ecosystem dependency — workloads that require RDS, EMR, SageMaker, Lambda, or other AWS-managed services without OpenMetal equivalents
- Global edge presence — AWS has 30+ regions; OpenMetal has four (Ashburn, Los Angeles, Amsterdam, Singapore)
- Short-term burst capacity — temporary high-memory capacity that does not justify a monthly commitment
Moving from AWS? Let’s Model Your Actual Costs.
Share your current AWS spend, egress volume, and storage requirements — we’ll provide a direct comparison against OpenMetal’s fixed-cost model.
Cost Model and TCO
| Factor | OpenMetal XXL v4 | AWS x2idn.32xlarge |
|---|---|---|
| Compute pricing | Fixed monthly | ~$26.93/hr on-demand |
| NVMe storage (38.4TB equivalent) | Included | 7.6TB ephemeral; EBS extra for persistence |
| Persistent storage (38.4TB EBS) | Included | ~$3,072/mo gp3 |
| Private network | Included (unmetered VLAN) | Cross-AZ: $0.01/GB |
| Public egress | 95th-percentile | $0.09/GB first 10 TB |
| IPMI / remote console | Included | Not available |
| DDoS protection | Included | Shield Standard (free) / Advanced ($3k/mo) |
| Commitment discount | Price lock up to 5 years | 1yr/3yr Reserved Instances |
12-Month TCO Estimate (sustained 24/7, 50 TB/month egress)
| Cost Item | OpenMetal XXL v4 | AWS x2idn.32xlarge |
|---|---|---|
| Compute (12 months) | Contact OpenMetal for quote | ~$235,500 [est. on-demand ~$26.93/hr] |
| Persistent storage (38.4TB EBS) | Included | ~$36,864 (38.4TB gp3 x 12 mo) |
| Egress (50 TB/mo x 12) | Contact OpenMetal for quote | ~$54,000 [est. at $0.09/GB] |
| Estimated total | Contact for quote | ~$326,000 |
The cost gap widens with egress volume. Workloads producing 100 TB/month of public egress add approximately $108,000/year in AWS transfer charges — a cost that does not scale linearly with OpenMetal’s 95th-percentile model. The persistent storage gap adds further: 38.4TB of EBS gp3 costs approximately $3,072/month with no analog on the OpenMetal XXL v4.
Deployment Options
OpenMetal XXL v4 is available in four data center locations — as a standalone bare metal server or as a three-node Hosted Private Cloud cluster. Ramp pricing is available for migrations from AWS.
| Location | Region | Certifications | Location Page |
|---|---|---|---|
| Ashburn, VA | US-East | SOC 1/2 Type II, ISO 27001, PCI DSS, NIST 800-53 HIGH, HIPAA (facility-level) | Ashburn |
| Los Angeles, CA | US-West | SOC 2, SOC 3, ISO 27001, PCI DSS | Los Angeles |
| Amsterdam | EU-West | SOC 1/2, PCI-DSS, ISO 27001, ISO 50001, ISO 22301 | Amsterdam |
| Singapore | Asia | BCA Green Mark Platinum | Singapore |
Get a XXL v4 Quote
Tell us about your infrastructure needs and we’ll provide a custom quote for the XXL v4.
- Bare metal: Single XXL v4 server, fixed monthly, 38.4TB persistent NVMe, IPMI included
- Hosted Private Cloud: 3-node OpenStack + Ceph cluster on XXL v4 hardware, Day 2 ops included
- Migration support: Ramp pricing for AWS-to-OpenMetal transitions
- Cost modeling: Share your current AWS spend and we’ll produce a direct comparison
Moving from AWS? We’ll model your actual usage — compute hours, egress volume, storage — against OpenMetal’s fixed-cost model.
Ramp pricing available for migrations. All deployments include fixed monthly pricing, 99.96%+ network SLA, and DDoS protection.
Specifications, pricing, and availability are subject to change without notice. The information on this page is provided for general guidance and does not constitute a contractual commitment. Contact OpenMetal for current configuration details and pricing. AWS specifications and pricing are sourced from publicly available documentation and may not reflect current rates or configurations.