GPU Cloud Provider · New Jersey, USA
Vultr
Vultr offers high-performance cloud GPU solutions globally, with extensive support for AI, ML, and HPC applications. They provide a cost-efficient, scale-ready cloud infrastructure featuring AMD and NVIDIA GPUs, simplified into user-friendly services with predictable pricing.
GPUs
5
Founded
2014
Countries
15
Data Centers
24
Uptime SLA
99.9%
Team Size
201-1000
GPU Marketplace
NVIDIA GB300 NVL72On-Demand
NVIDIA HGX B300On-Demand
NVIDIA L40S L40SOn-Demand
AMD Instinct MI300X MI300XOn-Demand
NVIDIA A100 80GB SXMOn-Demand
Company Profile
Company TypeScale-up
Provider TypeCloud Provider
Founded2014
HeadquartersNew Jersey, USA
Legal EntityVultr Holdings, LLC
FundingPrivate
Team Size201-1000
Infrastructure
GPU FleetAMD Instinct MI300X, AMD Instinct MI325X, NVIDIA A100 80GB, NVIDIA A16
Network FabricAdvanced CDN and network scalability with global coverage
ConnectivityData center-specific, ranging typically in high Gbps to Tbps, across a global scale
StorageBlock storage, File system storage, Object storage
Data Center TierTier 3
Bare MetalYes
AvailabilityGA (General Availability)
StartupEnterpriseDeveloperResearch
Compute & Deployment
On-DemandYes
Spot / InterruptibleNo
Reserved InstancesYes (monthly and longer-term committed pricing available)
Bare MetalYes (bare metal GPU servers available)
VM-BasedYes
Container-BasedYes (Docker)
KubernetesYes (managed Kubernetes via Vultr Kubernetes Engine)
Serverless GPUNo
Spin-Up Time2-5 minutes
TerraformYes (official provider on HashiCorp registry)
GPU Hardware
Latest GenH100 SXM, MI300X, MI325X
Legacy SupportA100 PCIe, A16
Multi-GPU NodesYes (up to 8x per node)
Max GPUs/Node8
NVLinkYes (NVLink on H100 SXM nodes)
InfiniBandYes (HDR 200Gbps on clustered nodes)
PCIe vs SXMBoth PCIe and SXM
HGX PlatformYes (HGX H100 8-GPU)
Pricing Model
Per HourYes (primary billing unit)
Per MinuteNo
SubscriptionNo
Spot DiscountNo spot pricing
Public PricingYes
Hidden FeesIP address charges apply; additional fees for managed databases and object storage
Egress ChargesFree up to 1TB/month per instance; $0.01/GB thereafter
Pay-as-you-goYes
Credit SystemYes (prepaid credits available)
Performance & Scaling
Multi-Node TrainingLimited (manual setup, no managed orchestration disclosed)
Elastic ScalingManual only
Auto ScalingNo
InfiniBandNo (Ethernet only, specific speeds not disclosed)
NVSwitchNo
SLA99.9%
Perf IsolationPartial (virtual GPU instances, not bare metal)
Noisy NeighborPartial (virtual isolation, multi-tenant infrastructure)
Developer Experience
OnboardingDeploy in under 5 minutes via web UI or API; no approval process required
FrameworksCUDA, cuDNN, ROCm
SDK LanguagesPython, Go, Node.js, PHP, Ruby, Java
CLI ToolingFull CLI (vultr-cli) for instance management, deployment, and automation
JupyterVia SSH port forwarding or user-configured Jupyter on deployed instance
TemplatesPyTorch, TensorFlow, CUDA Toolkit
DocumentationComprehensive docs with tutorials, API reference, and community guides
API FeaturesCLI, SDK, REST API, Terraform provider
Security & Compliance
SecurityRegular security assessments and compliance with key regulations such as GDPR and HIPAA
ComplianceGDPR, HIPAA
Founded 2014 with 10+ years of cloud infrastructure experience30+ global data center locationsPublicly documented 99.99% uptime SLASOC 2 complianceGDPR compliantPCI DSS compliant
Data Center Locations
Coverage
CountriesUnited States, Germany, Netherlands, United Kingdom, France, Sweden, Poland, Spain, Japan, South Korea, Singapore, Australia, India, Brazil, Canada
CitiesNew Jersey, Chicago, Dallas, Los Angeles, Seattle, Miami, Atlanta, Silicon Valley, Toronto, São Paulo, Amsterdam, Frankfurt, London, Paris, Stockholm, Warsaw, Madrid, Tokyo, Osaka, Seoul, Singapore, Sydney, Mumbai, Tel Aviv
Multi-Region FailoverYes (manual)
Latency TiersStandard cloud latency
North AmericaEuropeAsia-PacificSouth AmericaAfrica
Compliance Regions
EU Data ResidencyYes (Frankfurt, Amsterdam, London, Paris, Stockholm, Warsaw, Madrid)
US Gov CloudNo
India RegionYes (Mumbai)
Datacenter Locations
Key Strengths
AMD MI300X and MI325X availability — among first cloud providers to offer MI325X
Competitive per-hour pricing with no long-term commitment required
Global footprint across 30+ data center locations
Simple self-service onboarding with no sales process
Broad API and CLI support for infrastructure-as-code workflows
Known Limitations
GPU availability can be limited and subject to stock constraints
No spot/preemptible GPU pricing
Less mature AI/ML platform tooling compared to hyperscalers
No built-in model marketplace or managed ML pipelines
Limited enterprise-grade SLA customization for GPU workloads
Additional Information
Support Options
24/7 support through customer service portals and community resources
Community
Active community forums, Discord server, and GitHub presence; documentation community contributions accepted
Core Proposition
Globally distributed cloud infrastructure with straightforward pricing and broad GPU availability including high-end AMD Instinct accelerators for AI/ML workloads.
Payment Methods
Credit CardPayPalWire TransferCrypto (Bitcoin)
Last updated March 2026. Information subject to change.