Sovereign GPU & RDU Clusters
Fully managed 8×GPU nodes and RDU racks across Manchester & Düsseldorf. Canal-cooled, GDPR aligned, and ready for enterprise AI workloads.
H200 Cluster – 8×GPU
NVIDIA Hopper H200 compute nodes built for LLM training and sovereign-scale inference.
From £1.10 per GPU-hour
From £1,450 per GPU / month
- • 8×H200 GPUs (80 GB HBM3 each)
- • 960 GB total VRAM
- • Dual Xeon 8480+ CPUs
- • 400 Gb/s RoCE fabric – Canal-Cooled
£11,600
per 8×GPU cluster / month
B200 Cluster – 8×GPU
Next-gen NVIDIA Blackwell for hybrid precision AI workloads and generative inference.
From £2.50 per GPU-hour
From £2,150 per GPU / month
- • 8×B200 GPUs (192 GB HBM3e each)
- • 1.5 TB total VRAM
- • NVLink 5.0 architecture
- • Optimised for GenAI, Diffusion, Vision Transformers
£17,200
per 8×GPU cluster / month
SambaNova RDU Cluster
Full-rack SN40L system optimised for large-scale model inference and fine-tuning.
- • 8×RDU SN40L modules
- • 8 TB on-chip memory
- • AI compiler for LLM & transformer acceleration
- • Power: 6 kW / Phase – Sovereign-hosted
Ask for Quote
Tailored pricing for RDU workloads
Scale Your AI — Sustainably, Locally, Sovereignly.
UK and EU Data Centres • Canal-Cooled • 400 Gb/s Fibre Fabric • Renewable-Powered Compute Designed for AI Nations.