Hugging Face and Microsoft deepen collaboration: Azure open-model deployment, MI300X VMs, and Spaces Dev Mode

AI Impact Summary

Microsoft and Hugging Face broaden their strategic collaboration to simplify deployment of open models across Azure. The expanded Hugging Face Collection in the Azure Model Catalog enables one-click deployment from Azure AI Studio for models such as Llama 3, Mistral 7B, Command R Plus, Qwen 1.5 110B, and fine-tuned Open LLM Leaderboard entries, accelerating cloud adoption. The initiative also introduces AMD-based hardware acceleration with generally available ND MI300X VMs and Optimum-AMD/ROCm integration to boost inference performance, plus WebGPU and ONNX Runtime Web support for local browser inference and Phi-3 on-device demonstrations. Developer experience improvements include Spaces Dev Mode and VS Code integration, making it easier to string together Spaces demos with local development workflows.