Bare metal
Bare metal GPU servers built for value and scale
Take full control of your infrastructure with dedicated GPU servers engineered for performance, flexibility and demanding workloads.
Full control of your infrastructure
Dedicated physical machines give you the control, performance and predictability your workloads demand.
Infrastructure on your terms
Get consistent performance with dedicated infrastructure and network isolation.
Learn moreReady-to-go bare metal configurations
Choose from configurations built for speed, efficiency and real workload demands.
Host | Networking | Storage | Starting from | |
---|---|---|---|---|
8 × NVIDIA HGX SXM5 H100 80 GB Dual Intel Xeon 8480+ 56C (112T) @ 2GHz 2048 GB DDR5 ECC Memory | 800Gbps, 10Gbps SpectrumX, 100 Gbps upgradeable | NVMe SSD 2 × 1.9TB, 8 × 3.8TB, | from $23.12/hr Save up to $1.12/hr | |
8 × NVIDIA HGX PCIe H100 96 GB Dual Intel Xeon 8480+ 56C (112T) @ 2GHz 2048 GB DDR5 ECC Memory | 800Gbps, 10Gbps SpectrumX, 100 Gbps upgradeable | NVMe SSD 2 × 1.9TB8 × 3.8TB | from $21.52/hr | |
8 × NVIDIA A100 PCIe 80 GB Dual AMD Genoa 9454 48C (96T) @ 2.75GHz 2304 GB DDR5 ECC Memory | 800Gbps, 10Gbps SpectrumX, 100 Gbps upgradeable | NVMe SSD 2 × 1.9TB6 × 3.8TB | from $12.80/hr Save up to $0.80/hr | |
8 × NVIDIA L40S PCIe 48 GB Dual AMD Genoa 9454 48C (96T) @ 2.75GHz 2304 GB DDR5 ECC Memory | 800Gbps, 10Gbps SpectrumX, 100 Gbps upgradeable | NVMe SSD 2 × 1.9TB6 × 3.8TB | from $8.80/hr Save up to $0.40/hr | |
8 × NVIDIA HGX SXM5 H200 80 GB Dual Intel Xeon 8480+ 56C (112T) @ 2GHz 2048 GB DDR5 ECC Memory | 800Gbps, 10Gbps SpectrumX, 100 Gbps upgradeable | NVMe SSD 2 × 1.9TB8 × 3.8TB | from $23.29/hr | |
8 × NVIDIA B200 180 GB Dual Intel Xeon 8570 112C (224T) @ 4GHz 4096 GB DDR5 ECC Memory | 800Gbps, 10Gbps SpectrumX, 100 Gbps upgradeable | NVMe SSD 2 × 1.9TB8 × 3.84TB | Available on request | |
1 × NVIDIA GB200 NVL72 13.5 TB Arm® Neoverse V2 cores 2592C (2592T) @ 0GHz 17000 GB LPDDR5X Memory | 800Gbps, 10Gbps SpectrumX, 100 Gbps upgradeable | - | Available on request |
Trusted by NVIDIA. Built for your scale.
As a Preferred Partner for Compute, CUDO provides early access to NVIDIA’s most advanced hardware, from H100 and GB200 to the upcoming B300 and GB300. All systems are fully validated and tested across our infrastructure to ensure performance at scale.
- Be among the first to deploy B300 and GB300
- Run on certified, benchmarked infrastructure optimised for AI training and inference
- Deploy with confidence on systems built in alignment with hyperscaler standards
Also trusted by:
Leading partners across networking, storage and data centre operations, with a shared focus on performance and sustainability.
Scale with enterprise-ready infrastructure
Reserve GPUs, deploy clusters and scale on infrastructure designed for enterprise growth.
Frequently asked questions
Talk to sales
Enquire about bare metal access today, or reserve your bare metal on CUDO Compute for as long as you want it, with unique contracts tailored to suit your needs.
"CUDO Compute is a true pioneer in aggregating the world's cloud in a sustainable way, enabling service providers like us to integrate with ease"
VPS AI
Get started today or speak with an expert...
Available Mon-Fri 9am-5pm UK time