Hugging Face Hub integrates FriendliAI Endpoints for 1-click deployment on NVIDIA H100 GPUs

AI Impact Summary

Hugging Face is adding FriendliAI Endpoints as a deployment option in the Hub, enabling 1-click deployment of open-source or custom models to FriendliAI’s inference infrastructure. The integration leverages NVIDIA H100 GPUs with features like continuous batching, native quantization, and autoscaling to reduce latency and lower costs at scale. This shifts deployment and hosting to a managed service, reducing the operational burden for developers and enterprises while broadening access to high-performance inference. Expect faster time-to-value for model deployments and increased Hub adoption among users who prefer streamlined, cloud-hosted serving.