GPU Servers for AI

Rent GPU compute for local LLMs, AI agents, assistants, and demanding AI applications.

Cloud GPU

AWS (EC2)

Wide range of GPU instances in worldwide AWS infrastructure.

B200H200H100A100Trainium

Azure

GPU virtual machines integrated with Microsoft Azure services.

B200H200H100A100T4

CoreWeave

Specialized GPU cloud for enterprise AI and HPC workloads.

GB200$9.51/hB200$8.6/hH200$1.25/hH100$6.16/hA100$2.25/h

DataCrunch (Verda)

GPU cloud with A100/H100 for training and inference of AI models.

B200H200H100A100L40S

FluidStack

Flexible GPU cloud with competitive pricing for AI workloads.

GB200B200H200H100A100

Google Cloud (GCE)

Enterprise GPU instances integrated with Google Cloud ecosystem.

B200H200H100A100L4

Hyperstack

European GPU cloud with on-demand H100/A100 and per-minute billing.

H200$1.3/hH100$0.7/hA100$0.35/hL40$0.35/hRTX A6000

Jarvis Labs

GPU cloud optimized for AI model training and experiments.

H100$1.49/hA100$0.41/hRTX 6000RTX 5000A6000

Lambda Cloud

Cloud GPU servers optimized for deep learning and AI research.

B200$1.99/hH200$1.29/hH100$0.79/hA100$0.69/hGH200$1.29/h

OVHcloud

European cloud giant offering GPU instances in EU data centers.

H100$0.91/hA100$1/hL40S$0.91/hL4$0.6/hV100S$0.88/h

Paperspace (DigitalOcean)

Cloud GPU by DigitalOcean for ML, AI, and data science.

H100$0.76/hA100$0.76/hRTX 5000$0.51/hRTX 4000$0.45/hA6000$0.76/h

RunPod

On-demand and serverless GPU for ML/AI. Docker container support.

B200$2.09/hH200$1.79/hH100$1.59/hA100$1.09/hRTX 5090$0.49/h

Scaleway

French cloud with H100 and L40S instances in European data centers.

H100L40SL4RENDER-S

Together AI

GPU cloud for training and inference of AI models with API access.

B200$1.99/hH200$1.99/hH100$1.99/hA100

Vultr

Cloud GPU instances with global availability and simple API.

B200H200H100A100L40S

Peer-to-peer GPU cloud

Salad Cloud

Distributed GPU cloud with the lowest prices on the market.

RTX 5090$0.16/hRTX 4090$0.08/hRTX 3090$0.04/hRTX 3080$0.04/h

TensorDock

Affordable GPU servers with hourly billing.

H100$1.8/hA100$0.35/hRTX 5090RTX 4090$0.12/hA6000

Vast.ai

GPU catalog with the lowest prices. Peer-to-peer and cloud offerings.

H200H100A100RTX 5090RTX 4090

Dedicated GPU Servers

Latitude.sh

Dedicated bare metal GPU servers for demanding AI applications.

H200H100$1.68/hA100L40S

Data updated: 9/7/2026 · Prices are indicative, based on public price lists. The provider's official pricing applies.

What can you use GPU servers for?

Run local LLMs (Llama, Mistral, Qwen), train custom AI models, AI agent and assistant inference, image and video generation, scientific computing and simulations.