Stop opening 10 browser tabs to find the cheapest GPU.
GPU Tracker watches prices across 10 cloud providers — AWS, Google Cloud, Azure, Lambda Labs, RunPod, Vast.ai, CoreWeave, Paperspace, Hyperstack, and Salad — and shows you the cheapest current rate for every major NVIDIA GPU.
WHAT YOU CAN DO
- Browse live prices for H100, H200, A100 (40GB and 80GB), L40S, L4, A10, RTX 4090, RTX 3090, and B200
- Filter by GPU model, pricing type (on-demand, spot, community), and provider
- See a price history chart for any instance
- Set price alerts — for example, "notify me when 1x H100 spot drops below $1.50/hr"
- Get push notifications when an alert fires
- Tap any offer to jump straight to the provider's booking page
ASK AN AI ASSISTANT
Describe your workload — "fine-tune Llama 3 70B for 12 hours" or "cheapest A100 for an experiment" — and our AI assistant (powered by Anthropic's Claude) recommends the best option with reasoning. It knows the live price catalog and considers your duration, GPU count, and reliability needs.
WHO THIS IS FOR
- ML engineers and researchers training models on a budget
- Indie AI developers shopping for the best deal on inference
- Startup founders watching cloud burn rate
- Anyone curious where GPU prices are heading
HOW IT WORKS
We refresh prices every hour from each provider's public API or pricing page. Alerts evaluate every 15 minutes. The app is free to use.
PRIVACY
We store your email (for sign-in), your alerts, and your anonymous push token. That's it. No tracking, no ads, no selling your data. Full policy: sridharrajarao.com/privacy
ROADMAP
- Spot price feeds for AWS, GCP, Azure (currently estimated)
- Provider-direct booking via APIs
- Team accounts with shared alerts
Built by an indie developer. Questions or feedback: sridhar.rajarao@gmail.com
Live GPU prices from 10 clouds. Set alerts. Ask AI what to run.