About GPUniq — the GPU meta-cloud for AI workloads

GPUniq is a GPU meta-cloud: a rental marketplace that aggregates data-center and community-hosted GPUs into one platform with per-second billing, automatic failover, snapshots, and public real-time price analytics. We are the source of the pricing data we publish — not an aggregator of other people's numbers.

What we run

The marketplace tracks 70+ GPU models — from B200, H200 and H100 to RTX 5090/4090 consumer cards — across 30,000+ GPUs live in the network. Every model has a public statistics page with live minimum/median hourly prices, availability and 365-day history: no login, no paywall.

On top of the GPUs we operate an LLM API (OpenAI-compatible /v1/openai/* and native Anthropic /v1/messages) billed from the same balance, plus a CLI, an MCP server for AI agents, and a VRAM calculator.

Where our data comes from

Prices and availability on /statistics and the GPU Cloud Price Index are computed from live offers on our own marketplace and refresh every 5 minutes. Minimum is the cheapest rentable offer at measurement time; median is the median across all live offers for the model.

Company

GPUniq raised a pre-seed round in December 2025 and has been serving paying customers since early 2026. We are a small infrastructure team; you can talk to the founder directly — book a demo call.

Contact

GPU statisticsPrice IndexGPU guidesChangelog