Kinds of provider
How an AI compute provider makes its living.
Neoclouds
A neocloud is a cloud company built around GPUs for AI, rather than around general computing. It rents large numbers of Nvidia accelerators, joined by fast networks into clusters, to AI labs, start-ups and the big clouds themselves
Inference platforms
An inference platform runs trained AI models for customers and charges for the work, usually by the token, the small chunk of text a language model reads or writes. Customers send a request to an API and get an answer back, without renting a whole GPU
Sovereign and national AI clouds
A sovereign AI cloud is AI compute built so that a country, or a group of countries, controls it: the data stays at home, local law applies and the machines cannot be switched off from abroad. Governments fund some directly, back others through sovereign wealth funds, or subsidise local providers
Hyperscale clouds
The hyperscalers are the very large cloud companies: Amazon Web Services, Microsoft Azure and Google Cloud, followed by Oracle, Alibaba Cloud, Tencent Cloud, Huawei Cloud and IBM. They run data centres in dozens of countries and sell thousands of services
Developer clouds
Developer clouds sell simple, low-priced servers to programmers, small companies and start-ups, and most now add GPUs. DigitalOcean, Vultr, Linode (now Akamai Cloud), OVHcloud, Hetzner, Scaleway, Leaseweb and E2E Networks belong here
GPU marketplaces
GPU marketplaces match buyers of compute with many sellers instead of owning the hardware themselves. Some, such as Vast.ai and SF Compute, let data centres and owners list spare GPUs at prices that change with demand
Miners turned AI hosts
Some of the fastest-growing AI hosts began as bitcoin miners. Mining needs the same things as AI: cheap power, grid connections, land and cooling
Chipmaker clouds
A few AI chip companies sell their chips mostly as a service rather than as boxes. Cerebras, Groq and SambaNova each found that customers would rather pay for answers than learn to run unfamiliar hardware, so each built data centres and an API cloud on its own chips
AI supercomputers
An AI supercomputer is a single very large machine, or one cluster on one network, built to train big models or to run science at scale. Some belong to companies, such as xAI's Colossus, Meta's research clusters and Microsoft's Fairwater sites
Research and university clusters
Research clusters are the computers of universities, national labs and public research centres. They are paid for with public money and shared among scientists, usually by application rather than by payment