Snorkel AI and Hugging Face enable enterprise access to 150k+ foundation models via Inference Endpoints

AI Impact Summary

Snorkel Flow now leverages Hugging Face Inference Endpoints to broaden access to the Hugging Face model hub, giving users access to 150k+ open-source foundation models without hosting them onsite. The pause-and-resume capability enables per-client activation, slashing idle compute costs and supporting multi-tenant deployments. This shift moves inference to managed endpoints, introducing governance, security, and data routing considerations for model choice, versioning, latency, and compliance, while accelerating enterprise-ready AI workflows.