The AI developer cloud for GPU pods, serverless inference and clustering
Lightning AI (Grid.ai) records 10,313 monthly visits with -1.8% month-over-month growth and Usage-based / Contact for Pricing pricing. Thunder Compute records 102,147 monthly visits with -29% growth and Paid (per-minute instances, per-second sandboxes) pricing.

On-demand GPU cloud and enterprise AI infrastructure from the creators of PyTorch Lightning.
| Dimension / Metric | ![]() | ![]() |
|---|---|---|
| Official Domain | grid.ai | thundercompute.com |
| Pricing Model | Usage-based / Contact for Pricing | Paid (per-minute instances, per-second sandboxes) |
| Monthly Visits | 10.31K (10,313) | 102.15K (102,147) |
| MoM Traffic Delta | -192 (-1.8%) | -41.66K (-29%) |
| Primary Market & Share | 🇺🇸 United States (50.5%) | 🇺🇸 United States (53.5%) |
| Direct Visits / mo | 4.11K | 40.74K |
| User Rating | 4.4 | 4.4 |
| Supported Platforms | Website | Website |

On-demand GPU cloud and enterprise AI infrastructure from the creators of PyTorch Lightning.
Self-serve access to H200, B200, H100, A100 and L4 GPUs with guaranteed availability and no long-term contracts.
Code, debug, and refine models in the browser or over SSH on cloud GPUs, with shared studio environments for teams.
Run hundreds of experiments concurrently, compare runs, and scale to large multi-node GPU clusters for pretraining.
SSO, role-based access, audit logs, fine-grained data controls, SOC2 and HIPAA alignment, plus private cloud and VPC deployment.

Low-cost cloud GPUs with per-minute instances and per-second sandboxes for AI workloads.
Launch RTX A6000, L40, A100 80GB, or H100 SXM machines, often ready to SSH into within minutes, and pay only while they run.
Spin up isolated Ubuntu VMs with or without a GPU for RL rollouts, evals, and CI, billed per second and scaled to zero between runs.
A CLI, VS Code extension, MCP server, and REST API let you drive infrastructure from a terminal, an editor, an AI agent, or your own services.
Save an instance disk as a snapshot and restore it later, so you can pause experiments without leaving machines running.
The AI developer cloud for GPU pods, serverless inference and clustering
Rent cloud GPUs by the second from a global marketplace
Rent cloud GPUs by the second from a global marketplace
Free / Open-sourceServerless cloud infrastructure for AI and data workloads
Free / Open-sourceServerless cloud infrastructure for AI and data workloads
An AI-native cloud built from silicon to API for training, tuning, and serving models
An AI-native cloud built from silicon to API for training, tuning, and serving models

Free / Open-sourceDistributed GPU cloud that orchestrates the world's idle consumer GPUs
Free / Open-sourceDistributed GPU cloud that orchestrates the world's idle consumer GPUs
Free / Open-sourceThe AI cloud built by the creators of PyTorch Lightning
Free / Open-sourceThe AI cloud built by the creators of PyTorch Lightning
Free / Open-sourceOne cloud for 200+ model APIs, agent sandboxes and GPU compute

Gigawatt-scale AI compute, built and operated end to end

Agent-first GPU cloud: rent NVIDIA pods by the second from your terminal or API
Agent-first GPU cloud: rent NVIDIA pods by the second from your terminal or API

Private AI cloud for training and inference workloads
Private AI cloud for training and inference workloads

Free / Open-sourceServerless GPUs for real-time AI with sub-second cold starts
Free / Open-sourceServerless GPUs for real-time AI with sub-second cold starts

Rent cloud GPUs on demand for up to 80% less than hyperscalers
Rent cloud GPUs on demand for up to 80% less than hyperscalers