Microsoft Simplifies Windows AI Development with New Foundry Platform
Microsoft is empowering developers to build AI-powered applications on Windows with the launch of Windows AI Foundry. Announced at Build 2025, this platform expands upon the Windows Copilot Runtime service, offering a unified solution for local AI development.
Windows AI Foundry streamlines the process of fine-tuning, optimizing, and deploying AI models within Windows apps. This initiative aligns with Microsoft's focus on AI as a key growth area, despite the associated development costs.
Boosting the Windows AI Ecosystem
While Microsoft might not directly profit from third-party AI apps on Windows, the company aims to cultivate a thriving ecosystem by providing robust development resources. Windows AI Foundry helps achieve this by simplifying AI software creation on their flagship operating system.
The platform automatically detects a Windows machine's hardware and downloads the necessary software components for running specific AI models. It also ensures these components stay updated and provides tools to simplify model preparation.
Introducing Foundry Local: On-Device AI Power
Foundry Local, a new service within Windows AI Foundry, brings the power of AI models directly to client devices, including Windows and macOS. It offers modules for running AI models and tools on-device using the open-source ONNX Runtime, which accelerates AI performance across platforms.
Developers can use the command-line interface with prompts like "Foundry model list" and "Foundry model run" to manage and interact with local models. Foundry Local automatically detects device hardware (CPU, GPU, and NPU) and lists compatible models for developers.
An SDK is also available for easy integration of Foundry Local into apps. This empowers developers to leverage local AI capabilities within their applications seamlessly.
