Infrastructure
The infrastructure behind AI
Explore hardware, inference providers and the software that connects a model to a working application.

GPUs and hardware
Compare catalogue specifications and estimated model memory fit.
Cloud inference providers
Explore hosted models, recorded prices and provider profiles.
GPU useful life
Understand physical service, software support and operating economics.
GPU networking
Understand NVLink domains and the topology behind a cloud quote.
Software and compute stacks
Read about CUDA, alternative stacks and serving software.
Run a model locally
Guides for Ollama, MLX and GGUF with original sources.
Get help with a deployment
Discuss model selection, infrastructure requirements or GPU sourcing.
Start with your device
Use the device picker to find models within our memory estimate.