RunPodUnclaimed

Runpod is a cloud platform for running GPU workloads. Developers and AI teams rent GPUs on demand for three kinds of work: Pods for development and persistent compute, Serverless for API-based inference that scales with traffic, and Clusters for multi-node training and large batch jobs.

The platform offers more than 30 GPU models, from B200 and H100 cards to RTX 4090s, across 31 global regions. Users deploy their own containers, frameworks and code. Serverless queues requests and assigns them to workers, so teams do not have to build their own orchestration. The site says a Serverless endpoint costs nothing while it is not running.

Runpod is aimed at developers, researchers and companies building AI products. Customer case studies on the site describe teams moving off large cloud providers to cut infrastructure costs, and one describes scaling a live application from zero to over 1,000 requests per second.

Pricing is pay-as-you-go. Pods are listed per hour or per second, with Community Cloud and Secure Cloud options, and the lowest listed Pod price is $0.27 per hour for an RTX A5000. Clusters are listed from $2.15 per GPU per hour and are billed for what you use. Reserved clusters and enterprise agreements are quoted on request.

The site states that Runpod is SOC 2 Type II certified and GDPR and HIPAA compliant, and that Secure Cloud adds network isolation for workloads with stricter compliance needs. Teams can start a GPU environment from the site or talk to sales about reserved capacity.