State-of-the-art GPU infrastructure on a global scale

With Vultr, you can choose from an extensive range of virtual machines and bare metal servers, all equipped with a variety of high-performance GPUs, enabling you to accelerate a diverse range of workloads using only the resources you need.
Vultr Cloud GPU can scale from a fraction of a single card all the way up to multi-card clusters for the intensive workloads.

GPU NVIDIA GH200

NVIDIA GH200

Delivering breakthrough acceleration of large-scale AI, model training and inference, and high-performance computing (HPC) applications

Learn more
GPU Nvidia A100

NVIDIA HGX H100

Delivering unprecedented acceleration to power the world’s most advanced AI, data analytics, and HPC workloads

Learn more
GPU Nvidia HGX H100

NVIDIA A100

Enabling researchers and scientists to combine simulation, data analytics, and AI to drive scientific progress

Learn more

NVIDIA L40S

Combining powerful AI compute with best-in-class graphics and media acceleration to power the next generation of data center workloads

Learn more
GPU Nvidia A40

NVIDIA A40

Combining professional graphics with powerful compute and AI, to meet today’s design, creative, and scientific challenges

Learn more
GPU Nvidia A16

NVIDIA A16

Enabling virtual desktops and workstations with the power and performance to tackle any project from anywhere

Learn more

Large-scale dedicated clusters and flexible on-demand VMs, accelerated by NVIDIA GPUs

Cloud GPU, underpinned by the partnership of Vultr and NVIDIA, stands as a beacon for next-gen GPU- accelerated infrastructure. Sidestepping the usual complications of driver setups and licensing, it offers users a direct conduit to the raw power of NVIDIA GPUs for any computational endeavor.

Download the datasheet

no form fill or personal details required for access


Deploy faster with Vultr Container Registry

From HPC to conversational AI to medical imaging to recommender systems and more, Vultr Container Registry offers ready-to-use containers — in one place — to speed up your application development and deployment process. All powered by Vultr’s preconfigured GPU Stack of NVIDIA drivers, CUDA Toolkit, Docker, and more.

Vultr GPU Stack powers Cloud GPU

Vultr GPU Stack: a finely tuned and integrated operating system and software environment, which instantly provisions the full array of NVIDIA GPUs, pre-configured with NVIDIA CUDA Toolkit, NVIDIA cuDNN, NVIDIA Drivers, for immediate deployment.

Vultr Kubernetes Engine for Cloud GPU

Create GPU-accelerated Kubernetes clusters that will power your most resource-intensive workloads anywhere in the world. This powerful combination empowers developers and innovators to build sophisticated AI and machine learning systems that can handle even the most complex challenges. Learn about Vultr Kubernetes Engine

xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px" viewBox="0 0 512 432" enable-background="new 0 0 512 432" xml:space="preserve">

Get ready for Cloud GPU powered by the NVIDIA HGX H100, designed specifically for AI, machine learning, and HPC workloads in healthcare, finance, and beyond.

Backed by powerful, global infrastructure

Vultr's cloud platform is a highly robust and scalable infrastructure that empowers businesses with lightning-fast computing resources and unparalleled reliability.

Cloud Compute

Bare Metal

Managed Databases

Kubernetes Engine

Direct Connect Networking

Cloud Compute

Vultr Cloud Compute delivers a highly performant, scalable, and reliable cloud compute instances with global coverage for easily meeting the needs of any workload.

Learn more about Vultr Cloud Compute

Bare Metal

Vultr Bare Metal provides dedicated physical servers without the virtualization layer, giving you full control and access to the hardware resources.

Learn more about Vultr Bare Metal

Managed Databases

Vultr Managed Databases offer top-of-the-line cloud-based database solutions, ensuring security, high availability, and easy scalability for businesses of all sizes.

Learn more about Vultr Managed Databases

Kubernetes Engine

Vultr Kubernetes Engine is a fully managed container orchestration service that allows you to easily deploy, scale, and manage containerized applications without the hassle of managing the underlying infrastructure.

Learn more about Vultr Kubernetes Engine

Direct Connect Networking

Vultr Direct Connect allows customers to establish a private network connection between their own infrastructure and Vultr's cloud data center locations.

Learn more about Vultr Direct Connect

Easy to manage

Our easy-to-use control panel and API let you spend more time coding and less time managing your infrastructure.

Vultr Control Panel
Vultr Control Panel - Current charges
Vultr Control Panel - Current charges
Vultr Control Panel - Dashboard on desktop
Vultr Control Panel - Choose Application
Vultr Control Panel - Choose server Location
Vultr Control Panel - Available Addons
Vultr Control Panel - One-Click Deploy
Vultr Control Panel - Backup Status
Vultr Control Panel - DDoS Status
Vultr Control Panel - Server Actions
Vultr Control Panel - Manage Server

Powerful platform worldwide

100% SLA guaranteed

Enterprise grade stability and performance.

API & Terraform

Use our API or Terraform provider to quickly create and control your instances.

Best-in-class AMD & Intel CPUs

The latest generation AMD and Intel CPUs guarantee fast performance.

NVMe SSD

Blazing fast read/write speeds on NVMe solid-state disk hardware.

Virtual private clouds

Establish isolated regional VPCs for your applications.

Public IPv6 network

Develop for the next generation internet protocol.

Root administrator access

Full "root" access and a dedicated IP address included with all VMs.

Custom ISOs

Deploy any operating system of your choosing.

Additional resources

Documentation and API access for cloud compute.

Documentation

API

FAQ

How am I billed for my instances?

All servers on your account are billed hourly up to the monthly rate cap. The hourly rate is determined by dividing the monthly rate by 672 hours (28 days). If your server is online for more than 672 hours in a calendar month, you will only be billed the monthly rate. Accumulated charges are invoiced to your account on the 1st of every month.

If you upgrade your instance to a larger plan, the instance will appear on separate invoice line items in the month you upgraded. Each line item will bill at the hourly rate associated with the plan for the actual number of hours during that period of time. Because of this, you may be billed more than 672 hours (up to the actual number of hours in that month) during a month that you have upgraded your instance.

Do you charge for stopped instances?

Yes, instances in a stopped state continue to reserve dedicated system resources (RAM, SSD storage, IP aliases, vCPU) and therefore incur charges until you destroy the instance. If you wish to no longer accumulate charges for a virtual machine, please use the DESTROY button in the customer portal.

What operating system templates do you offer?

See our full list of operating systems.

Does Vultr collect VAT/Sales tax?

Vultr.com is required to collect tax in several countries around the world. This tax is generally referred to as VAT (Value Added Tax), consumption tax, or sales tax. Prices listed on our website do not include tax. Tax will be added to your invoice as a separate line item.

What is the bandwidth overage rate?

We charge $0.01 per GB for bandwidth used in excess of your quota.

View all FAQs

How am I billed for my instances?

All servers on your account are billed hourly up to the monthly rate cap. The hourly rate is determined by dividing the monthly rate by 672 hours (28 days). If your server is online for more than 672 hours in a calendar month, you will only be billed the monthly rate. Accumulated charges are invoiced to your account on the 1st of every month.

If you upgrade your instance to a larger plan, the instance will appear on separate invoice line items in the month you upgraded. Each line item will bill at the hourly rate associated with the plan for the actual number of hours during that period of time. Because of this, you may be billed more than 672 hours (up to the actual number of hours in that month) during a month that you have upgraded your instance.