12 companies tracked in this category. Names and one-line descriptions are free to browse — pricing model, funding, founding year, headquarters, and cited sources are in the full dataset.
| Company | Description |
|---|---|
| CoreWeave | CoreWeave provides GPU cloud infrastructure, storage, and managed services on a Kubernetes-native platform purpose-built for AI training and inference workloads. |
| Crusoe Energy | Crusoe designs and operates vertically integrated, energy-focused AI data centers and a cloud platform providing NVIDIA and AMD GPU compute for AI training and inference. |
| FluidStack | FluidStack operates a cloud platform that provisions large-scale NVIDIA GPU clusters on demand for AI training and inference, and builds/operates dedicated AI data centers for large AI labs. |
| Hyperstack (NexGen Cloud) | Hyperstack, owned by NexGen Cloud, is a self-serve GPU-as-a-Service cloud platform offering instant access to NVIDIA GPU virtual machines for AI, ML, and rendering workloads. |
| Lambda | Lambda provides GPU cloud computing and AI infrastructure, including on-demand instances and large multi-GPU clusters, for AI training and inference workloads. |
| Modal | Serverless cloud platform that lets developers run GPU-accelerated AI/ML training, inference, and general compute workloads with automatic scaling. |
| Nebius | Publicly traded AI infrastructure company providing GPU cloud computing and data-center capacity for AI training and inference, formed from the restructuring of Yandex N.V. |
| RunPod | RunPod is a cloud platform offering on-demand and serverless GPU compute for developing, training, and deploying AI applications with per-second billing. |
| SF Compute | Operates an online exchange/spot market for large-scale GPU cluster capacity, letting buyers purchase or resell GPU-cluster time by the hour or minute. |
| Together AI | Cloud platform offering GPU clusters, serverless inference APIs, and fine-tuning tools for running and building generative AI and open-source models. |
| Vast.ai | Vast.ai runs a peer-to-peer marketplace connecting individuals and data centers with spare GPUs to renters seeking low-cost, on-demand or interruptible compute for AI/ML workloads. |
| Voltage Park | Operates NVIDIA GPU cloud infrastructure offering on-demand and reserved rental access for AI training and inference; merged with Lightning AI in January 2026 to form a combined AI cloud. |