News
European technologies in security and space
Commissioner Andrius Kubilius' visit to CloudFerro highlights the importance of European technologies for Europe's security and operational readiness
Public Cloud · Compute
GPU Services
On-demand GPU compute in EU data centres, from RTX-class acceleration to H200 clusters up to 32 GPUs. Built for AI, ML, HPC, and rendering workloads with integrated cloud networking and storage.
- EU Sovereign & GDPR Compliant
- Up to 32×H200 clusters
- vGPU · Passthrough · Cluster
Features
Scalable GPUs for advanced compute workloads
CloudFerro GPU Services provide immediate access to modern NVIDIA GPUs across three delivery models.
Broad NVIDIA GPU portfolio
RTX A6000, RTX PRO 6000 Blackwell, L40S, A100, H100, and H200, available as shared vGPU VMs, dedicated PCI passthrough instances, and standard multi-node H200 clusters. Availability varies by region and delivery model.
Three delivery models
Shared vGPU VMs for rapid experimentation and cost efficiency; dedicated PCI passthrough instances for predictable near-native performance; standard H200 multi-node clusters up to 32 GPUs for distributed training and HPC.
Data performance & storage integration
Local NVMe scratch where applicable, plus fully integrated block and object storage for dataset staging, checkpointing, and high-throughput data access across the full AI and HPC pipeline.
Self-service provisioning via portal & APIs
Provision, manage, and automate GPU resources through the CloudFerro portal or REST APIs. Supports IaC workflows and automation tooling for repeatable, scalable deployments.
Scale compute power without investing in your own hardware
Our GPU VMs offer unparalleled performance for demanding tasks like training extensive AI models, conducting large-scale inference, and handling vast datasets with ease
Accelerated Training
Reduce model training time from weeks to hours. Scale training across multiple GPUs for even faster results.
Cost Efficiency
Pay only for what you use with hourly billing. No upfront hardware investment or maintenance costs.
Enterprise Support
Technical support with SLA guarantees. Dedicated account managers for enterprise customers.
Data Proximity
Process Earth observation data with GPUs located in the same data center. Eliminate data transfer bottlenecks.
Typical Use Cases
From model training to HPC workloads
INTEGRATIONS
Works natively with the CloudFerro ecosystem
GPU instances integrate with CloudFerro storage, networking, and managed AI services — enabling end-to-end AI and HPC pipelines without moving data outside the EU.
Block Storage
Attach persistent SSD volumes to GPU instances for high-throughput data access, model checkpointing, and output storage. Data persists independently of instance lifecycle.
Object Storage
Stage large training datasets and store model artefacts with S3-compatible object storage, directly accessible from GPU instances within the same sovereign environment.
Virtual Networks
Full OpenStack networking: private networks, routers, floating IPs, and security groups, enabling isolated GPU workloads and secure multi-tenant operation.
Managed GenAI Services (Sherlock)
Complement GPU compute with CloudFerro managed GenAI inference, combine dedicated training infrastructure with a privacy-first LLM API for full-stack AI pipelines.
APIs & IaC automation
Provision and manage GPU resources via REST APIs or OpenStack Heat orchestration templates. Fully compatible with Terraform, Ansible, and common AI/HPC automation tooling.
AI/ML frameworks
Compatible with TensorFlow, PyTorch, CUDA, RAPIDS, and other major ML and HPC frameworks, deployable on demand with standard NVIDIA driver support across all GPU models.
Scope description
Technical Specification
GPU-accelerated compute in EU data centres, delivered as shared vGPU VMs, dedicated PCI passthrough instances, and standard H200 clusters — integrated with CloudFerro cloud networking and storage for end-to-end AI and HPC pipelines.
GPU portfolio
RTX A6000 · RTX PRO 6000 Blackwell · L40S · A100 · H100 · H200 (availability by region and model)
Delivery models
Shared vGPU VMs · Dedicated PCI passthrough · Standard H200 clusters up to 32 GPUs
Storage
Local NVMe scratch (where applicable) · Integrated block and object storage
Provisioning
Self-service via portal and REST APIs · IaC / automation compatible
Networking
Full OpenStack · Private networks · Security groups · Floating IPs
Pricing model
Usage-based and custom · Pricing tables available per cloud region
FAQ
Frequently Asked Questions
Didn’t find the answer you were looking for? Get in touch if you need more information or would like to speak with our team.
Other Products
Compute
Virtual Machines
Run and scale public cloud workloads in EU-based OpenStack data centres. Configure CPU, RAM, storage and OS, manage lifecycle actions.
Storage
Block Storage
Create and manage persistent block storage volumes, backups, snapshots and volume groups.
Storage
Object Storage
Manage object storage containers and buckets for S3/Swift-compatible data access.
Containers & Serverless
Managed Kubernetes
Deploy production-grade Kubernetes clusters through GUI, CLI or Terraform without managing the control plane yourself.
Databases
Managed Databases
Managed instances of PostgreSQL, MySQL, MariaDB, Valkey, and MS SQL Server. Automated backups, failover, and monitoring.
AI & Machine Learning
Managed GenAI Services
Use a fully managed, privacy-first GenAI service with OpenAI-compatible endpoints, hosted on 100% European infrastructure with processing in Poland.