Vultr features: tier-by-tier list (2026)
Bare-metal-friendly cloud with cheap shared-CPU starter VMs and dedicated AMD EPYC tiers. Bandwidth caps and add-ons drive the real bill.
Core capabilities (38)
AMD Instinct™ MI355X
Providing exceptional acceleration with breakthrough memory capacity and memory bandwidth.
Source · core
Nvidia A16
Enabling virtual desktops and workstations with the power and performance to tackle any project from anywhere.
Source · core
Kubernetes
Deploy and scale containerized apps with a fully managed service. Vultr Kubernetes Engine ushers in a better way to optimize container orchestration.
Source · core
Optimized Cloud Compute
Powered by the latest-generation AMD and Intel CPUs, spin up virtual machines for general purpose or optimized configurations in under 60 seconds.
Source · core
Managed Databases
Secure, highly available, and easily scalable. Vultr Managed Databases for MySQL, PostgreSQL, Apache Kafka®, and Valkey™ just work right out of the box.
Source · core
Bare Metal
Stay in full control of your environment with high performance single-tenant dedicated servers, accelerated by NVIDIA GPUs and high-performance CPUs from AMD and Intel.
Source · core
Develop locally, deploy globally ®
33 Global Data Center Regions Amsterdam NL Test Speed Atlanta , GA US Test Speed Bangalore IN Test Speed Chicago , IL US Test Speed Dallas , TX US Test Speed Delhi NCR IN Test Speed Frankfurt DE Test Speed Honolulu , HI US Test Speed Johannesburg ZA Test Speed
Source · core
Wireguard By vultr-labs
View All “Athos is committed to providing precisiontherapeutics for patients with autoimmune diseases and cancer.
Source · core
Harness a wide variety of plug-and-play SaaS applications
Harness a wide variety of plug-and-play SaaS applications to make developing and deploying your cloud applications easier.
Source · core
“Vultr has been more than just a cloud
“Vultr has been more than just a cloud provider for us – they’ve been a true innovation partner.
Source · core
Operations
[ edit ] Vultr provides access to cloud infrastructure services and virtualized GPUs. [ 14 ] [ 15 ] The company operates data centers across North America, South America, Europe, Asia, the Middle East, Sub-Saharan Africa, and Australia. [ 1
Source · core
Categories
VX1™ VMs optimized for affordable core infrastructure workloads. Cloud Compute Easy-to-use, affordable VMs for many common workloads. Optimized Cloud Compute No noisy neighbors on these powerful VMs with built-in NVMe SSD.
Source · core
General Purpose
These virtual machines offer a typical balance of CPU and RAM resources suitable for common workloads. vCPUs Memory Bandwidth Storage Hourly Price 2 vCPUs 8 GB Memory 5.00 TB Bandwidth Block Storage Storage $0.060 /hr 4 vCPUs 16 GB Memory 6.00 TB Bandwidth Blo
Source · core
Memory Optimized
These virtual machines offer proportionately more RAM than CPU resources for workloads with high memory requirements. vCPUs Memory Bandwidth Storage Hourly Price 2 vCPUs 16 GB Memory 5.00 TB Bandwidth Block Storage Storage $0.080 /hr 4 vCPUs 32 GB Memory 6.00
Source · core
Cloud Compute
These virtual machines run atop shared vCPUs, and are suitable for many business and personal applications: low traffic websites, blogs, CMS, dev/test environments, small databases, and much more.
Source · core
Regular Performance
Powered by previous generation Intel CPUs and regular SSD. vCPUs Memory Bandwidth Storage Monthly Price Hourly Price 1 vCPU 0.5 GB Memory 0.50 TB Bandwidth 10 GB Storage $2.50 /mo $0.004 /hr IPv6 Only 1 vCPU 0.5 GB Memory 0.50 TB Bandwidth 10 GB Storage $3.50
Source · core
AMD Pricing
vCPUs Memory Bandwidth Storage Monthly Price Hourly Price 1 vCPU 1 GB Memory 2.00 TB Bandwidth 25 GB Storage $6.00 /mo $0.009 /hr 1 vCPU 2 GB Memory 3.00 TB Bandwidth 50 GB Storage $12.00 /mo $0.018 /hr 2 vCPUs 2 GB Memory 4.00 TB Bandwidth 60 GB Storage $18.0
Source · core
CPU Optimized
For compute bound applications, these VMs provide proportionally more CPU than they do RAM and NVMe SSD.
Source · core
Storage Optimized
These VMs provide especially generous amounts of NVMe SSD storage to accompany a typical balance of CPU and RAM. Use cases: large nonrelational databases e.g.
Source · core
Amd Mi355x
Providing remarkable acceleration with breakthrough memory capacity and memory bandwidth. The prices shown in the table below reflect on-demand preemptible instances for AMD MI355X GPUs. 48-month prepaid pricing starts at $2.650/GPU/hr.
Source · core
Direct Connect
Privately connect Vultr to other cloud service providers, colocation facilities, or on-premises data centers.
Source · core
Databases
Databases that 'just work' right out of the box. Quickly deploy a new database cluster while letting us manage the configuration, updates, and end-to-end security, so you can stay focused on your app.
Source · core
PostgreSQL
PostgreSQL - This open source object-relational database system has more than 30 years of active development and a reputation for reliability, feature robustness, and performance.
Source · core
Vultr Managed Databases for Valkey™
Vultr Managed Databases for Valkey is a fully managed in-memory NoSQL data store that is purpose-built for large amounts of consistently changing data. Vultr Managed Databases for Valkey can be used as a database, cache, streaming engine, and message broker.
Source · core
Block Storage
NVMe: High-performance storage for workloads requiring rapid I/O. HDD: Globally available, affordable, and scalable persistent hard disk storage.
Source · core
Object Storage
Reliable S3-compatible object storage for varied performance needs. Select each performance tier for pricing details.
Source · core
File System (NVMe)
High-performance, scalable, and affordable file storage. Price: $ 1.00 /month Storage: 10 GB High Performance Starting from $1/10GB
Source · core
Storage Gateways
Connect Vultr Bare Metal servers to Vultr File System instances quickly and easily through Vultr Storage Gateways. For high availability and additional throughput, Highly Available Vultr Storage Gateways are coming soon.
Source · core
NAT Gateways
Vultr NAT Gateways provide Network Address Translation (NAT) and internet connectivity for instances within a Vultr VPC, including VPC-Only Instances.
Source · core
Load Balancers
Deploy fully configured and managed load balancers rapidly for seamless load distribution.
Source · core
Serverless Inference
Intelligently deploy and serve GenAI models without the complexity of infrastructure management.
Source · core
Choose High Performance or High Frequency plans for
Choose High Performance or High Frequency plans for the newer generations of AMD or Intel CPUs, along with NVMe SSD.
Source · core
These virtual machines run atop fully dedicated, new
These virtual machines run atop fully dedicated, new generation AMD EPYC vCPUs. Dedicated vCPUs ensure that these machines deliver the fast, consistent performance that many business applications require.
Source · core
Often a good starting point, these VMs provide
Often a good starting point, these VMs provide a typical balance of CPU, RAM, and NVMe SSD resources.
Source · core
Use cases: open source databases e.g
Use cases: open source databases e.g. MySQL, in-memory databases and caches (e.g.
Source · core
Use cases: large nonrelational databases e.g
Use cases: large nonrelational databases e.g. Cassandra and MongoDB, high frequency online transaction processing (OLTP)
Source · core
Delivering up to 10X higher performance for applications
Delivering up to 10X higher performance for applications running terabytes of data, enabling scientists and researchers to reach unprecedented solutions for the world’s most complex problems.
Source · core
When using Vultr Kubernetes Engine, you only pay
When using Vultr Kubernetes Engine, you only pay for underlying worker nodes, as well as associated resources, e.g.
Source · core
AI features (17)
AMD Instinct™ MI325X
Setting new standards in HPC, AI training, and deployments with inference-optimized acceleration and efficiency.
Source · ai
AMD Instinct™ MI300X
Delivering powerful and efficient AI and HPC deployments with exceptional performance and inference-optimized acceleration.
Source · ai
Nvidia Gb300 Nvl72
Delivering unprecedented performance for the age of AI reasoning with NVIDIA Blackwell Ultra GPUs and NVIDIA Grace™ CPUs in a liquid-cooled, rack-scale architecture.
Source · ai
Nvidia Hgx™ B300
Accelerating the era of AI reasoning with NVIDIA Blackwell Ultra GPUs’ exceptional memory capacity and memory bandwidth.
Source · ai
Nvidia Hgx™ B200
Delivering a new era of accelerating computing and generative AI, integrating NVIDIA Blackwell GPUs with high-speed interconnect to accelerate AI performance at scale.
Source · ai
Nvidia Gh200
Delivering breakthrough acceleration of large-scale AI, model training and inference, and high-performance computing (HPC) applications.
Source · ai
Nvidia Hgx™ H100
Delivering unprecedented acceleration to power the world’s most advanced AI, data analytics, and HPC workloads.
Source · ai
NVIDIA HGX™ A100 and A100 PCIe
Enabling researchers and scientists to combine simulation, data analytics, and AI to drive scientific progress.
Source · ai
Nvidia A40
Combining professional graphics with powerful compute and AI, to meet today’s design, creative, and scientific challenges.
Source · ai
Nvidia Hgx B300
Supercharging AI and HPC with exceptional memory capacity and memory bandwidth. Contact sales to reserve capacity now!
Source · ai
Nvidia Hgx B200
Propelling a new era of accelerating computing and generative AI with NVIDIA Blackwell GPUs for the most demanding AI, high-performance computing (HPC), and data analytics workloads. The prices shown in the table below reflect 48-month contract pricing.
Source · ai
Nvidia Hgx H100
Designed specifically for AI and HPC workloads, featuring fourth-generation Tensor Cores and the Transformer Engine with FP8 precision for accelerated performance. The prices in the table below reflect 24-month contract pricing.
Source · ai
Nvidia Hgx A100
Delivering unprecedented acceleration and powering the world's highest-performing AI, data analytics, and HPC workloads. Contains eight A100 80 GB SXM GPUs with 600 Gbps NVLink interconnect, optimized for workloads requiring multiple GPU cards.
Source · ai
Amd Mi325x
AI Training and AI Inference 8 x AMD MI325X 256 GB 2 x 960 GB NVMe 8 x 3.84 TB NVMe 2 x AMD EPYC 9575F 128 cores / 256 threads @ 3.3 GHz 3 TB RAM 15 TB Bandwidth 8 x 400 Gbps GPU Network 100 Gbps Network Starting from $2.000 /GPU/hour
Source · ai
NVIDIA A100 PCle
AI Training and AI Inference 4 x NVIDIA A100 80 GB PCle 2 x 480 GB SSD 6 x 1.6 TB NVMe 2 x Intel Gold 6248R 48 cores / 96 threads @ 2.8GHz 512 GB RAM 15 TB Bandwidth 25 Gbps Network Starting from $1.290 /GPU/hour
Source · ai
Chat / Prompt Models
Model Price per Million Input Tokens Price per Million Output Tokens nvidia/Llama-3.1-Nemotron-Safety-Guard-8B-v3 $0.01 /M tokens 0.01 $0.01 /M tokens nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 $0.13 /M tokens 0.38 $0.38 /M tokens nvidia/Nemotron-Casca
Source · ai
Image Generation Models
Image generation is only available through the API. Model Price per Megapixel stable-diffusion-3.5-medium 0.01 $0.01 /MP flux.1-schnell 0.01 $0.01 /MP stable-diffusion-3.5-large-turbo 0.02 $0.02 /MP flux.1-dev 0.02 $0.02 /MP
Source · ai api
Integrations (1)
Use cases: web and application servers, E-Commerce, game
Use cases: web and application servers, E-Commerce, game servers, video and audio streaming, API serving, relational databases
Source · integrations api
Collaboration (1)
Vultr Managed Apache Kafka®
Apache Kafka® - the trusted highly-available, open source real-time data stream processor. Stream data with low latency globally and flexibly.
Source · collab real-time
Reporting and analytics (2)
This website uses first and third party cookies
This website uses first and third party cookies for advertising purposes, to better understand your preferences, to offer you an optimal user experience, to keep track of statistics and to be able to collect your feedback.
Source · reporting
Use cases: video encoding, batch processing, CI/CD, high
Use cases: video encoding, batch processing, CI/CD, high performance computing (HPC), ad serving, analytics processing
Source · reporting
Administration (2)
Cloud GPU
Vultr makes accelerated computing affordable and easy to use. In just minutes, provision VMs or bare metal servers on demand.
Source · admin
¹Each NVIDIA GB300 NVL72 is sold as a
¹Each NVIDIA GB300 NVL72 is sold as a single continuous NVLink domain and contains 18 of the compute nodes above for a total of 72 GPUs.
Source · admin
See Vultr pricing in detail
This page lists what Vultr does. For per-seat math, tier comparison, and the full hidden-fee list, see the Vultr pricing breakdown.
Closest alternatives
Other cloud iaas tools buyers shortlist when they evaluate Vultr.
Sources
Features evolve. Confirm directly with the vendor before purchase decisions. This list was last refreshed on 2026-05-27.
Frequently asked questions
What features does Vultr include?
This page tracks 61 verified Vultr features across 6 categories, sourced from Vultr's official site and Wikipedia.
Which tier do I need for Vultr?
Vultr's feature availability varies by plan. Compare the public Vultr tiers, calculator, and hidden fees on the Vultr pricing page.
How does Vultr compare on features to alternatives?
Buyers commonly compare Vultr against DigitalOcean, Linode, Hetzner Cloud. Each alternative's feature page is linked below where we cover it.