Microsoft Unveils Windows AI Foundry to Empower Developers at Build 2025
Microsoft introduces Windows AI Foundry at Build 2025, providing a unified platform to streamline AI model selection, optimization, and deployment across Windows devices. This evolution of the Windows Copilot Runtime aims to make Windows the premier open platform for AI development.
Microsoft introduces the Windows AI Foundry at Build 2025, an evolution of the Windows Copilot Runtime designed to support the entire AI developer lifecycle. This unified platform allows developers to seamlessly handle model selection, optimization, fine-tuning, and deployment across both client devices and the cloud. The company emphasizes its commitment to keeping Windows an open and flexible platform tailored for the new era of AI development.
At the core of this announcement is Windows ML, which serves as the built-in AI inferencing runtime for the operating system. Developers bring their own models and deploy them efficiently across a wide silicon partner ecosystem that includes AMD, Intel, NVIDIA, and Qualcomm. The platform spans CPUs, GPUs, and NPUs to ensure broad hardware compatibility for various AI workloads.
The Windows AI Foundry also integrates directly with Foundry Local and popular model catalogs like Ollama and NVIDIA NIMs. This integration gives developers quick access to ready-to-use open-source models, enabling them to easily browse, test, interact with, and deploy models in their local applications. Additionally, ready-to-use AI APIs powered by inbox models on Copilot+ PCs handle key language and vision tasks to further streamline the development process.