RunPod

RunPod, founded in October 2022 by Runpod Inc., is a cloud computing platform providing on-demand GPU infrastructure for AI workloads, letting users train models, run inference, and execute batch processing without managing physical hardware, serving more than 1 million developers at companies including Hugging Face, Perplexity, and Replit. Pods offer GPU instances as Reserved (guaranteed) or Spot (lower-cost, interruptible) options across 31 global regions; Serverless provides autoscaling endpoints with sub-200ms cold starts via FlashBoot technology, scaling from zero to thousands of workers with zero idle cost; and Clusters support multi-GPU distributed computing for large-scale training and batch inference. RunPod offers persistent network storage without egress fees, and enterprise reliability with a 99.9% uptime SLA, SOC 2 Type II compliance, and HIPAA and ISO 27001 certifications available. Pricing is pay-as-you-go by the hour or second with no contracts or minimum commitments, and Serverless endpoints incur charges only while actively running.