GPU Cloud Provider
Hostkey
Hostkey is a dedicated server and colocation provider with roots in Europe, offering high-performance bare metal infrastructure including GPU-equipped servers for compute-intensive workloads. The company focuses on providing affordable dedicated hardware with flexible configurations for developers and enterprises.
GPUs
3
Countries
2
Data Centers
3
Team Size
51-200
GPU Marketplace
Company Profile
Company TypeScale-up
Provider TypeBaremetal Provider
HeadquartersNetherlands
FundingPrivate
Team Size51-200
Infrastructure
GPU FleetAMD Radeon RX 7900 XTX, AMD Radeon RX 6900 XT, AMD Radeon Pro W6800, NVIDIA RTX 4090, NVIDIA RTX 3090, NVIDIA A100, NVIDIA RTX A5000
Data Center TierTier 3
Bare MetalYes – all GPU offerings are bare metal dedicated servers
EnterpriseStartupResearch
Compute & Deployment
On-DemandYes (monthly billing, deploy immediately)
Spot / InterruptibleNo
Reserved InstancesYes (monthly and longer-term contracts available)
Bare MetalYes
VM-BasedNo
Container-BasedYes (Docker)
KubernetesNo
Serverless GPUNo
Spin-Up TimeHours (manual provisioning and setup)
TerraformNo
GPU Hardware
Latest GenRadeon RX 7900 XTX, Radeon Pro W7800
Legacy SupportRadeon RX 6900 XT, Radeon Pro W6800
Multi-GPU NodesYes (up to 2x per node on select configurations)
Max GPUs/Node2
NVLinkNo (AMD Radeon GPUs do not support NVLink)
InfiniBandNo
PCIe vs SXMPCIe only
HGX PlatformNo (HGX is NVIDIA-specific; AMD Radeon nodes use standard PCIe)
Pricing Model
Per HourYes (primary billing unit)
Per MinuteNo
SubscriptionYes (monthly plans available)
Spot DiscountNo spot pricing
Public PricingPartial (some GPUs listed)
Pay-as-you-goYes
Performance & Scaling
Multi-Node TrainingLimited (manual setup required, no managed distributed training orchestration)
Elastic ScalingNo (static dedicated server allocations)
Auto ScalingNo
NVSwitchNo (PCIe-based servers, no SXM form factor for Radeon)
Perf IsolationYes (dedicated bare metal servers, no virtualization layer)
Noisy NeighborYes (bare metal, no sharing with other tenants)
Developer Experience
OnboardingOrder via web UI; provisioning typically within hours for in-stock configurations
CLI ToolingWeb console and IPMI/KVM remote management; SSH access standard
JupyterVia SSH port forwarding
DocumentationBasic documentation with hardware specs and setup guides; limited tutorials
Data Center Locations
Coverage
CountriesNetherlands, Russia
CitiesAmsterdam, Moscow, Saint Petersburg
EuropeNorth America
Compliance Regions
EU Data ResidencyYes (Amsterdam, Netherlands)
US Gov CloudNo
India RegionNo
Datacenter Locations
Key Strengths
Bare metal dedicated GPU servers with no virtualization overhead
AMD Radeon GPU availability rare among cloud providers
Competitive pricing on high-end consumer and prosumer GPU hardware
European data center presence with GDPR-friendly infrastructure
Flexible hardware customization options
Known Limitations
Primarily monthly billing – no per-hour or spot pricing
Limited orchestration and managed ML tooling compared to hyperscalers
No native container orchestration or Kubernetes managed service
Smaller GPU fleet and fewer GPU SKUs than major cloud providers
Limited public documentation and tutorials
Additional Information
Core Proposition
Dedicated bare metal GPU servers with AMD Radeon and NVIDIA options, offering full hardware isolation without virtualization overhead.
Payment Methods
Credit CardWire TransferPayPalCrypto
Last updated March 2026. Information subject to change.