GPU Marketplace
Estimates only — rates are collected automatically from public provider pages and may be out of date.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 TypeGPU marketplace
Legal EntityLightning AI
Infrastructure
GPU FleetH100
Total GPU Capacity40,000 GPUs
Bare Metaluse managed Kubernetes, SLURM, bare-metal, or Lightning AI’s orchestrator
AI teams
Compute & Deployment
On-DemandYes
Spot / InterruptibleYes
Reserved InstancesYes
Bare MetalYes
Container-BasedYes
KubernetesYes
Spin-Up Timeself-serve in minutes
GPU Hardware
Latest GenOn-demand H100s
Multi-GPU NodesTrain models from scratch on large IB GPU clusters with large datasets.
Pool Size40,000 GPUs
InfiniBandlarge IB GPU clusters
Pricing Model
Per HourFrom $4.50/hr.
SubscriptionNo long-term contracts.
Public PricingCheck out our pricing page for the latest prices across clouds, or see them in the Lightning platform when requesting a machine.
Egress ChargesLightning provides cloud agnostic folders that let you store data that can easily move across clouds without ingress/egress fees.
Pay-as-you-goWhen you start a workload (like a Studio or Job) without reserving machines in advance, it runs on-demand. You pay only for the time you use.
Credit SystemStart now with 80 free GPU hours
Performance & Scaling
Multi-Node TrainingTrain models from scratch on large IB GPU clusters with large datasets.
Elastic ScalingElastic multi-cloud portability
InfiniBandTrain models from scratch on large IB GPU clusters with large datasets.
Developer Experience
OnboardingStart now with 80 free GPU hours; Guaranteed availability, self-serve in minutes.
Documentationquick tutorial; Deploy Examples (full code); various docs pages (Build, Inference, Train, GPU Cloud, Bring your own cloud)
Security & Compliance
SOC2HIPAAGDPRTrusted by 450000+ devs and AI teamsnamed customer testimonials
Data Center Locations
Coverage
Compliance Regions
Datacenter Locations
Key Strengths
Unified multi-cloud
flexible compute choices including spot/on-demand/reserved
supports managed Kubernetes/SLURM/bare-metal
no vendor lock-in
expanded provider network for availability
Known Limitations
On-demand availability depends on supply and demand (popular GPU types may be temporarily sold out)
reservations require contacting support
Additional Information
Core Proposition
Stop renting AI. Own it instead.
Notable Customers
Infineon
NNAISENSE
SLALOM
INWORLD
Columbia University
RETICULAR AI
CLEATUS
Last updated August 2026. Information subject to change.