RunPod

RunPod was founded in 2022 to provide affordable cloud GPU compute for machine learning workloads. The platform offers two main products: Pods (container-based persistent or interruptible GPU instances with NVIDIA H100, A100, RTX 4090, and other accelerators) and Serverless (pay-per-second autoscaling inference endpoints).

Users deploy custom Docker images or community templates and can rent GPUs by the hour or commit to discounted reserved capacity. Workloads run across data centers in North America and Europe via a mix of owned infrastructure and partner colocations.

RunPod is bootstrapped and grew to serve thousands of ML developers and small AI teams competing on price and developer experience with hyperscaler GPU offerings.