Neoclouds
CoreWeave, Nebius, Lambda, Crusoe, Together and Nscale: clouds built around GPUs for AI. 101 providers on EveryNeocloud so far.
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. The word spread in 2024 as CoreWeave, Nebius, Lambda, Crusoe and Nscale grew from niche hosts into companies with multi-billion-dollar contracts.
How a neocloud differs
A general cloud such as AWS sells hundreds of services, from databases to email. A neocloud sells mostly one thing: time on GPU servers, either by the hour or in reserved blocks of a year or more. It tunes the whole stack for that job, with dense liquid-cooled racks, InfiniBand or fast Ethernet between servers, shared storage close to the GPUs and software for scheduling training runs. Nvidia supports the model through its Nvidia Cloud Partner programme and has invested in several neoclouds.
Who buys, and the risks
The largest buyers are the AI labs and the hyperscalers. Microsoft accounted for 62 percent of CoreWeave's 2024 revenue, and Microsoft, Meta and OpenAI have signed large contracts with CoreWeave, Nebius, Nscale and IREN. Supporters say neoclouds build faster than the giants. Critics point to heavy borrowing secured on GPUs, to dependence on a few customers and to the risk that chips lose value quickly as new generations arrive.