Hugging Face NVIDIA NIM API (serverless) deprecated — migrate to Inference Providers

AI Impact Summary

The Hugging Face NVIDIA NIM API (serverless) service, built in collaboration with NVIDIA, is deprecated as of April 10, 2025. This service offered a serverless way to run inference on popular Generative AI models like Llama and Mistral using NVIDIA DGX Cloud accelerated compute, simplifying infrastructure management and reducing upfront costs. Developers can now leverage standardized APIs and a few lines of code within the Hugging Face Hub to access state-of-the-art open Generative AI models, but this service is no longer available.