Hugging Face Models on Foundry Managed Compute

Microsoft Foundry is a platform for building AI applications that now offers a curated catalog of open-weight models from the Hugging Face ecosystem, refreshed weekly and deployable with one click onto Foundry Managed Compute. Foundry Managed Compute is a managed GPU service for open-source and custom models where Microsoft handles infrastructure maintenance while developers manage model configuration. This matters because open-weight models have closed the gap with proprietary models on benchmarks while offering advantages like deep customization, hosting control, cost flexibility, and version control, but the operational challenges of security screening, runtime selection, and enterprise-grade deployment have historically made them difficult to use at scale. The curation pipeline systematically identifies trending models, screens them for compliance and security, selects appropriate runtimes, and stages weights in Azure so enterprises get the same security, governance, and observability features for open models as for other models on the platform.