Definition
Neocloud is a neocloud is a cloud company built mainly to rent out GPU computing power for AI training and inference, rather than offering the broad catalog of services a giant like Amazon or Google does.
A neocloud is a cloud provider whose main product is access to GPUs for AI work. You rent time on racks of chips, usually by the hour, month or multi-year contract, and run your own training or inference jobs on them. CoreWeave, Lambda, Crusoe and Nebius are commonly named examples.
Think of it as a specialty shop next to a big-box store. The large general clouds, often called hyperscalers, sell hundreds of services: storage, databases, email, networking. A neocloud sells one thing, dense clusters of AI chips, and tunes everything around keeping them busy.
Neoclouds buy chips, mostly from Nvidia, and put them in data centers they build or lease. Then they rent that capacity to AI labs, startups and companies that cannot get enough GPUs elsewhere. Because chips are expensive, many neoclouds finance their purchases with debt and sign long contracts with a few large customers to cover the payments.
That setup is the reason people watch them closely. A neocloud earns money only while its chips are rented at a price that covers the loan, and the chips lose value as newer generations arrive.
Most people never rent a neocloud directly. You meet one indirectly, when the AI service you use runs on rented capacity, or when your company trains or hosts a model on a small GPU provider to save money or get chips sooner.
If that describes you, a few practical points follow:
A hyperscaler offers GPUs too, alongside everything else, and may be easier if you already run your other systems there. A neocloud may be quicker to get chips from or cheaper for a pure AI workload, but it offers fewer surrounding services. Many teams use both.
Neoclouds exist because AI needs enormous compute, and the chips that supply it, GPUs and the high-bandwidth memory beside them, are scarce. Software built on CUDA runs on most of them, which is part of why Nvidia hardware dominates these clouds.
A neocloud is a cloud provider focused on renting GPU computing power for AI, instead of a broad catalog of cloud services.
A hyperscaler such as a big general cloud sells many kinds of services, while a neocloud concentrates on dense GPU clusters for AI workloads.
Elon Musk said on X that SpaceX closed $40 billion in debt to buy Nvidia GPUs and predicted most of the 200-plus neoclouds will die or merge. Bloomberg reported pricing talks were still going. If you rent GPUs from a small cloud, read your contract.