RunPod
On-demand serverless GPU cloud (A100, H100, RTX series) with autoscaling and per-second billing. The go-to choice for indie AI developers and teams that need GPU compute without committing to AWS or GCP reserved instances.
Lambda Labs
Lambda Labs provides on-demand GPU cloud instances for model training and a serverless inference API for popular open-source models. With competitive pricing and high-end H100/A100 availability, it's a go-to for teams that need both training compute and inference.