GPU Cloud Provider · United Kingdom
Civo
Civo provides high-performance NVIDIA GPUs optimized for machine learning, generative AI, and Kubernetes workloads, offering flexible, scalable infrastructure solutions. With a focus on cost-effectiveness, safety, and environmental sustainability, their cloud GPU services cater to a wide array of computational needs and ensure data sovereignty and security compliance in the UK.
GPUs
2
Founded
Unknown
GPU Marketplace
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
Estimates only — rates are collected automatically from public provider pages and may be out of date. Prices vary by region, commitment term, and availability, and typically exclude storage, egress, and tax. Confirm current pricing with the provider before purchasing. Last collected .
Company Profile
Company TypeTransparent Cloud Pricing
HeadquartersUnited Kingdom
Legal EntityCivo
Infrastructure
GPU FleetNVIDIA L40S 48GB, NVIDIA A100 40GB, NVIDIA A100 80GB, NVIDIA H100 SXM (80GB), NVIDIA H200 SXM (141GB HBM3e), NVIDIA B200 (180GB)
Network FabricState-of-the-art networking capable of handling complex multi-GPU setups
ConnectivityData not specified
StorageHigh-performance storage solutions suited for intensive computational workloads (specific types not detailed)
AvailabilityLive
AI/ML training and inference, rendering, and scientific workloads
Compute & Deployment
On-DemandOn-demand
Reserved InstancesCommitment - 6 months Commitment - 12 months Commitment - 24 months Commitment - 36 months
VM-Basedhigh-performance virtual machines
Container-BasedCivo Kubernetes
KubernetesGPU Clusters on Civo Kubernetes
Spin-Up Timeunder 60 seconds
TerraformTerraform
GPU Hardware
Multi-GPU NodesCivo CUDA image works on both single-GPU and multi-GPU instances
NVLinkNVLink fabric is relevant on H100 hosts; single-GPU H100 hosts require NVreg_NvLinkDisable=1 to avoid NVLink initialisation, and NVLink is left enabled when an NVLink switch is present on multi-GPU hosts
PCIe vs SXMThe document explicitly labels H100 and H200 GPUs as SXM variants
Pricing Model
Per HourWe charge by the hour
SubscriptionWe provide our services on a straightforward monthly billing cycle.
Reserved DiscountWe can also offer preferential rates and discounts as part of a revenue commitment agreement.
Public Pricingcivo.com/pricing
Hidden FeesNo caveats, No fineprint. No surprise bills.
Egress ChargesFree & unlimited
Pay-as-you-goYou simply pay for the resources you use, allowing you to budget your cloud expenses and not be caught out with a big bill of products you don't need.
Credit SystemSign up and get $250 free credits
Performance & Scaling
Multi-Node TrainingIf you want to run GPU workloads on Kubernetes instead of a single instance, see GPU Clusters on Civo Kubernetes.
NVSwitchThe Civo CUDA image detects whether an NVLink Switch device is present on the host and disables NVLink on single-GPU hosts.
Developer Experience
OnboardingSign up and get $250 free credits
FrameworksSupports major ML frameworks conducive for Kubernetes integrations (specific frameworks not mentioned)
CLI ToolingCivo CLI with commands such as 'civo region list' and 'civo region use'; image and instance creation also possible via Dashboard and Terraform.
TemplatesCivo CUDA-enabled Ubuntu disk images (examples: ubuntu-cuda13-1, listed as 24.04-cudaXX-Y)
DocumentationGuides and how-tos for GPU instances including methods to obtain GPU-ready instances, region availability guidance, disk image selection, CLI commands, step-by-step instance creation flow, and NVLink troubleshooting instructions.
API FeaturesDetailed interaction likely via cloud management APIs compatible with Kubernetes setups (specific API details not mentioned)
Security & Compliance
SecurityISO 27001 certification; specifics on regular security practices such as pentesting not detailed
ComplianceCE, CE+, ISO 27001
Trusted by top-tier teams worldwideProvides Dashboard, Civo CLI and Terraform integrations
Data Center Locations
Coverage
worldwide
Compliance Regions
Datacenter Locations
Key Strengths
Simple and predictable pricing
Claims up to 75% less costly than hyperscale cloud providers
Billing model described as cloud-optimised
Free core control plane components for Civo Kubernetes (only pay for worker nodes and paid addons)
Free & unlimited data transfer
Fast VM launches ("launching in under 60 seconds")
Multiple NVIDIA GPU types available and CUDA-ready disk images provided
Known Limitations
GPU compute is not available in every Civo region
Single-GPU H100 hosts require disabling NVLink (NVreg_NvLinkDisable=1) or using the Civo CUDA image to avoid NVIDIA driver/fabric initialization failures
Additional Information
Support Options
Slack Connect channel, Direct access to SRE team
Core Proposition
Simple and predictable pricing
Last updated March 2026. Information subject to change.