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
1
Founded
2014
Countries
20
Data Centers
33
GPU Marketplace
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Estimates only — rates are collected automatically from public provider pages and may be out of date.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
Provider TypeCloud GPU
Founded2014
HeadquartersNew Jersey, USA
Legal EntityVultr
Infrastructure
GPU FleetNVIDIA GB300 NVL72, AMD Instinct™ MI455X, NVIDIA HGX™ B300, AMD Instinct™ MI355X, NVIDIA HGX™ B200, AMD Instinct™ MI325X & MI300X, NVIDIA GH200 Grace Hopper™ Superchip, NVIDIA HGX™ H100, NVIDIA HGX™ A100 & A100 PCIe GPU, NVIDIA L40S GPU, NVIDIA A40 GPU, NVIDIA A16 GPU
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
Bare MetalVultr Bare Metal provides direct access to the full power of the physical hardware, without a virtualization layer.
AvailabilityGA (General Availability)
AIMLgaming3D renderingAI model trainingmachine learning & deep learningvideo rendering & 3D modelingHigh-performance gaming & game development
Compute & Deployment
On-DemandPricing shown is on-demand and billed hourly.
Reserved InstancesReserved GPU servers on Vultr offer consistent, full-performance access to the entire GPU
Bare MetalVultr Bare Metal provides direct access to the full power of the physical hardware; Bare Metal servers are always single-tenant
VM-BasedCloud Compute: Easy-to-use, affordable VMs for many common workloads.
Container-BasedVultr Container Registry and containerized services in Vultr Marketplace; Vultr Kubernetes Engine to create GPU-accelerated Kubernetes clusters
KubernetesVultr Kubernetes Engine for Cloud GPU: Create GPU-accelerated Kubernetes clusters
Serverless GPUVultr Serverless Inference automatically provisions, scales, and shuts down GPU resources based on real-time demand
Spin-Up Timeinstant provisioning in under 15 seconds
TerraformUse our API or Terraform provider to quickly create and control your instances.
GPU Hardware
Latest GenNVIDIA GB300 NVL72, AMD Instinct™ MI455X, NVIDIA HGX™ B300, AMD Instinct™ MI355X, NVIDIA HGX™ B200, AMD Instinct™ MI325X, AMD Instinct™ MI300X, NVIDIA GH200 Grace Hopper™ Superchip, NVIDIA HGX™ H100, NVIDIA HGX™ A100, NVIDIA A100 PCIe, NVIDIA L40S GPU, NVIDIA A40 GPU, NVIDIA A16 GPU
Multi-GPU NodesVultr Clusters supports deploying and managing clusters of Vultr Cloud GPUs
NVLinkNVLink support is indicated for A100 SXM configurations (600 Gbps NVLink interconnect)
PCIe vs SXMBoth PCIe and SXM A100 variants are offered
HGX PlatformNVIDIA HGX™ B300,NVIDIA HGX™ B200,NVIDIA HGX™ H100,NVIDIA HGX™ A100
Pricing Model
Per HourPricing shown is on-demand and billed hourly.
SubscriptionOn-demand (billed hourly); pre-paid contract terms; pay-as-you-go
Public Pricinghttps://www.vultr.com/pricing/
Egress ChargesVultr states it delivers services without ... high egress costs.
Pay-as-you-gotransparent pay-as-you-go rates and no long-term contracts required.
Performance & Scaling
Multi-Node TrainingWith Vultr Clusters, deploy and manage clusters of Vultr Cloud GPUs without reservations or manual requests.
Elastic ScalingScale and configure clusters at will to meet varying workload demands or to optimize performance.
Auto ScalingVultr Serverless Inference automatically provisions, scales, and shuts down GPU resources based on real-time demand.
Perf IsolationBare Metal servers are always single-tenant, which means that only you have access to the resources of your server.
Noisy NeighborNo noisy neighbors on these powerful VMs with built-in NVMe SSD.
Developer Experience
OnboardingSelf-service, on-demand Vultr Clusters available through the Vultr Console or API; sign up for a free Vultr account.
FrameworksCUDA, cuDNN, ROCm
SDK LanguagesAPI and Terraform provider
TemplatesPre-configured AI/ML templates, plug-and-play SaaS applications in Vultr Marketplace, pre-built Kubernetes containers in Vultr Container Registry.
Model MarketplaceVultr Marketplace and Vultr Container Registry (marketplace of pre-built Kubernetes containers).
DocumentationFAQ and Vultr Cloud GPU Doc; API and Terraform provider; Datasheet.
API FeaturesCLI, SDK, REST API, Terraform provider
Security & Compliance
SecurityRegular security assessments and compliance with key regulations such as GDPR and HIPAA
ComplianceGDPR, HIPAA
AMD is part of the Vultr Cloud AllianceVultr is an NVIDIA Partner Network (NPN) Preferred Cloud Partnercustomer case study quote from Captions.
Data Center Locations
Coverage
CountriesNL, US, IN, DE, ZA, GB, AU, MX, ES, IT, JP, FR, CL, BR, KR, SG, SE, IL, CA, PL
CitiesAmsterdam NL, Atlanta, GA US, Bangalore IN, Chicago, IL US, Dallas, TX US, Delhi NCR IN, Frankfurt DE, Honolulu, HI US, Johannesburg ZA, London GB, Los Angeles, CA US, Madrid ES, Manchester GB, Melbourne AU, Mexico City MX, Miami, FL US, Milan IT, Mumbai IN, New Jersey, NJ US, Osaka JP, Paris FR, Santiago CL, São Paulo BR, Seattle, WA US, Seoul KR, Silicon Valley, CA US, Singapore SG, Stockholm SE, Sydney AU, Tel Aviv IL, Tokyo JP, Toronto CA, Warsaw PL
Amsterdam NLAtlanta, GA USBangalore INChicago, IL USDallas, TX USDelhi NCR INFrankfurt DEHonolulu, HI USJohannesburg ZALondon GBLos Angeles, CA USMadrid ESManchester GBMelbourne AUMexico City MXMiami, FL USMilan ITMumbai INNew Jersey, NJ USOsaka JPParis FRSantiago CLSão Paulo BRSeattle, WA USSeoul KRSilicon Valley, CA USSingapore SGStockholm SESydney AUTel Aviv ILTokyo JPToronto CAWarsaw PL
Compliance Regions
India RegionBangalore IN,Delhi NCR IN,Mumbai IN
Datacenter Locations
Key Strengths
Self-service on-demand GPU clusters via Console or API
bare metal and VM GPU offerings
partnerships with AMD and NVIDIA
global presence across 33 regions
transparent on-demand hourly pricing
Serverless Inference for auto-scaling GenAI.
Known Limitations
No spot instance pricing information provided
no uptime SLA stated
no company funding/employee/founded details
no PUE or green energy commitment details provided.
Additional Information
Support Options
24/7 support through customer service portals and community resources
Core Proposition
Vultr makes accelerated computing affordable and easy to use; simplifies next-gen GPU-accelerated infrastructure and sidesteps driver setups and licensing
Notable Customers
Captions
Last updated March 2026. Information subject to change.