GPU Cloud Provider · Luxembourg
Gcore
Gcore is a European cloud provider specializing in AI-driven solutions, notably through its Generative AI Cluster utilizing NVIDIA A100 and H100 GPUs. It offers robust cloud services aimed at accelerating AI/ML model development and providing high-performance, scalable infrastructure to various sectors including healthcare, finance, and gaming.
Founded
Not specified
Company Profile
Provider TypeHigh-performance AI GPU cloud solution
FoundedNot specified
HeadquartersLuxembourg
Legal EntityGcore
Infrastructure
GPU FleetL40S, A100, H100, H200, B300, GB300
Network FabricInfiniBand
ConnectivityNetwork capacity over 110 Tbps
StorageS3/NFS
Bare MetalBare Metal GPU | Dedicated physical servers with guaranteed resources. No virtualization overhead
AvailabilityNot specified
GamingMedia & StreamingEcommerce & RetailERP HostingFintechHosting ProvidersEcommerce AgenciesOnline EducationiGaming
Compute & Deployment
On-DemandFlexible on‑demand and reserved pricing
Spot / InterruptibleSpot Bare Metal GPU (interruptible with a 24-hour notice)
Reserved InstancesReserved instances available
Bare MetalBare Metal GPU (dedicated physical servers with guaranteed resources, no virtualization overhead)
VM-BasedVirtual GPU / Virtual Machines available
Container-BasedSupports containerized workloads using Docker and Kubernetes
KubernetesManaged Kubernetes (containerized deployments for GPU applications)
Serverless GPUNo - GPU Cloud provides dedicated infrastructure (not serverless)
Spin-Up Time15–40 minutes
TerraformYes - Terraform integration
GPU Hardware
Latest GenNVIDIA H100, H200, and upcoming GB200
Multi-GPU NodesMulti‑GPU cluster support
Max GPUs/Node8
NVLink72-GPU NVLink domain with Grace-Blackwell Ultra superchips
InfiniBandInstances are available with or without InfiniBand support
PCIe vs SXM8× B300 SXM with 2.3 TB VRAM
Pricing Model
Subscriptionon-demand, reserved pricing, and flexible billing options
Spot Discountup to 50% discount
Public Pricing/pricing/ai
Egress Chargesfree
Pay-as-you-gopay‑as‑you‑go
Performance & Scaling
Multi-Node TrainingYes, our GPU Cloud supports multi-GPU configurations and distributed training for large-scale AI workloads.
Max Cluster SizeClusters can scale to hundreds of nodes. Production deployments with 250+ nodes in a single cluster are supported, limited only by regional stock availability.
Elastic ScalingResize: Add or remove nodes based on demand.
Auto ScalingDynamically scale GPU clusters based on real-time workload demands, ensuring optimal efficiency.
InfiniBandInfiniBand is configured automatically when you create a cluster. If the selected configuration includes InfiniBand network cards, all nodes are placed in the same InfiniBand domain with no manual setup required.
Perf IsolationDedicated physical servers with guaranteed resources. No virtualization overhead
Developer Experience
OnboardingSign up for a Gcore account, select your desired GPU configuration, and deploy your instance through our user-friendly dashboard.
FrameworksNot specified
JupyterJupyter
Templatespre-configured environments and containerized workloads using Docker and Kubernetes
DocumentationProduct docs; API reference; developer tools; guides for creating Bare Metal and Virtual GPU clusters; complete documentation index available
API FeaturesCLI, SDK, REST API, Terraform provider
Security & Compliance
SecurityCompliance with PCI DSS, ISO/EIC 27001, GDPR
ComplianceGDPR, PCI DSS, ISO/EIC 27001
Customer case studies (MicrosoftProSiebenHiggsfieldFawkes Games)Gcore status page for uptime and incident trackingenterprise-grade security and DDoS protection24/7 support
Data Center Locations
Coverage
Globally available
Compliance Regions
Datacenter Locations
Key Strengths
Optimized for AI/ML
Bare metal and virtual GPU options
Automated API and Terraform control
Intelligent auto-scaling
Ultra-fast InfiniBand networking
Multi-GPU cluster support
Flexible on-demand and reserved pricing
Global availability and security features
Known Limitations
Cluster configuration cannot be changed after creation
local NVMe storage is temporary (deleted with cluster)
GB300 pricing is 'Per request' (no public rate listed)
Additional Information
Support Options
24/7 technical support
Community
Discord; Blog; Events; Startup Program
Core Proposition
Accelerate AI training, inference, and high‑performance computing with cutting‑edge NVIDIA GPUs and ultra‑low latency network
Notable Customers
Microsoft
ProSieben
Higgsfield
Fawkes Games
Last updated March 2026. Information subject to change.