← back to the directory
Hosted Service Inference & Serving

Vast.ai

1Marketplace 2for 3renting 4cheap 5spare 6GPUs

the six Ws · specification

W1 Who

Individuals and startups on a budget who need cheap on demand GPU compute for training or inference.

W2 What

Vast.ai is a peer to peer GPU cloud marketplace where hosts list spare GPU capacity and renters bid for compute by the hour.

W3 Where

Fully hosted, accessed through a web console, CLI, Python SDK, or REST API, running workloads on distributed host machines worldwide.

W4 When

Operating since 2018 and continuously expanding its host network through 2026.

W5 Why

Delivers substantially lower prices than traditional clouds by matching renters directly with individual and data center GPU hosts.

W6 With

Provides an official vast-cli Python package and Docker based instance templates for common ML frameworks.

W7 Watch

Commercial marketplace requiring an account and payment, dynamic per hour pricing, workloads run on third party host hardware vetted by Vast.ai.

gpu-marketplacespot-pricingclidockerpeer-to-peer

for agents & scripts

Reading this as a machine? Query it directly.

Search is open JSON - no key. Report telemetry after using a tool and it feeds that tool’s Proof Score. Or speak MCP to /mcp and discover tools mid-loop.