Vast.ai
GlobalGPU marketplace with flexible access to consumer and datacenter GPUs.
T4RTX PRO 6000A100H100H200B200
Starting from$0.44/hrCatalog price
View provider →
gpufleetContact →MANAGED GPU INFRASTRUCTURE
Your GPUs.
Your cloud.
We manage infrastructure.
Find GPU capacity through trusted providers. We deploy and manage the Kubernetes GPU platform behind it, so your team can run production workloads without the operational complexity.
12-MONTH GPU PRICE INDEX
See where GPU prices are heading.
Compare 3 months, 6 months, 1 year, or all retained history using the median daily market low. Current data refreshes the latest point.
Prices are recorded daily. Each monthly point is the median of its daily market lows. Recorded history does not expire and is available under All. Current prices are indicative and are not a quote.
GPU CLOUD MARKETPLACE
Compare GPU providers.
Compare hardware, pricing and availability across specialized GPU cloud providers.
SPOT
RunPod
GlobalOn-demand GPU cloud for AI, machine learning and container workloads.
RTX PRO 6000A100H100H200B200
Starting from$0.69/hrCatalog price
View provider →Lambda
US / GlobalGPU cloud infrastructure designed for AI development and training.
A100H100H200B200B300
Starting from$1.29/hrCatalog price
View provider →CoreWeave
US / EuropeSpecialized cloud infrastructure built for large-scale GPU workloads.
RTX PRO 6000A100H100H200
Starting from$1.49/hrCatalog price
View provider →Nebius
Europe / USAI-focused cloud infrastructure with high-performance GPU compute.
RTX PRO 6000H100H200B200B300
Starting from$1.25/hrCatalog price
View provider →Crusoe
USGPU cloud infrastructure for AI training, inference and compute.
A100H100H200B200
Starting from$1.35/hrCatalog price
View provider →TensorDock
GlobalDistributed GPU cloud providing flexible and cost-efficient compute.
T4A100H100
Starting from$0.42/hrCatalog price
View provider →Vultr
GlobalCloud GPU infrastructure for AI training, inference and accelerated computing.
A100H100
Provider pricingView pricingCheck provider
View provider →AWS
GlobalHyperscale cloud GPU instances for AI training, inference and enterprise workloads.
T4A100H100H200B200B300
Starting from$1.86/hrCatalog price
View provider →Microsoft Azure
GlobalEnterprise cloud GPU virtual machines for AI, HPC and accelerated computing.
T4A100H100H200
Starting from$3.40/hrCatalog price
View provider →Google Cloud
GlobalCloud GPU infrastructure for model development, training and scalable inference.
T4A100H100H200B200
Starting from$0.56/hrCatalog price
View provider →DigitalOcean
North AmericaDeveloper-focused GPU Droplets for AI development, inference and training.
H100H200B300
Starting from$1.57/hrCatalog price
View provider →Some provider links may be referral links. gpufleet may earn a commission at no additional cost to you. GPU pricing and availability may change.
GPU MANAGEMENT
Configure your GPU infrastructure.
Estimate the cost of GPU capacity and the Kubernetes operations needed to run it, including partitioning, scheduling, and observability.
4
Estimated GPU and management costESTIMATE
Estimated hourly$8.64/hr
Estimated monthly$6,311
GPU compute$7.96/hr
Management$400/mo
Add-ons$100/mo
Estimates are based on configured GPU and gpufleet management options. Actual provider pricing may vary.
PRODUCTS
Expert guidance and operations for your GPUs.
GPU Consultancy
Get expert advice on GPU selection, cluster architecture, workload performance, and infrastructure costs.
MIG Configuration
Partition supported NVIDIA GPUs into isolated instances with dedicated compute and memory for Kubernetes workloads.
GPU Time-Slicing
Share GPU access across development, testing, and intermittent workloads to make better use of available capacity.
GPU Observability
Track GPU utilization, memory, and health with dashboards and alerts that help your team troubleshoot and plan capacity.
GPU CONSULTANCY
Make the right decisions for your GPUs and GPU cluster.
Planning a new cluster or improving the GPUs you already have? We help you assess your workloads, choose suitable infrastructure, and build a practical plan for performance, reliability, and cost.
Explore GPU consultancy →GPU selectionWorkload fit
Cluster architectureCapacity planning
PerformanceBottleneck review
Infrastructure costsOptimization plan
gpufleet PRICING
Simple infrastructure management.
GPU compute costs depend on the provider and hardware you select. gpufleet focuses on helping you find, deploy and manage the right infrastructure.
GPU provider pricingCompare & choose
Compare providers →CONTACT
Ready to build your GPU infrastructure?
Tell us what you are running and we can help you choose the right GPU infrastructure.
Contact gpufleet →