At its annual Build 2025 conference, Microsoft announced Windows AI Foundry, a new platform that simplifies the creation of AI-powered applications for the Windows operating system. This platform is an expanded and updated version of the previous Windows Copilot Runtime service.
Microsoft describes the new platform as “a single ecosystem for local AI development.” It allows developers to build, optimize, and run their AI models on Windows-based devices. In the meantime, it’s clear that the company sees AI as one of its strategic strengths — Microsoft CEO Satya Nadella announced in January that the company’s AI revenue has reached $13 billion per year.
What’s changed?
Windows AI Foundry significantly automates the development environment:
- Automatically detects device hardware (CPU, GPU, NPU) and recommends suitable AI models;
- Automatically downloads and updates the necessary components;
- Simplifies the process of building, testing, and integrating AI models.
Windows AI Foundry also includes a new Foundry Local service. This tool is designed to run models directly on a user’s device. Foundry Local:
- Runs on Windows and macOS;
- Built on the ONNX Runtime;
- Supports running models via the command line
- Makes it easier to integrate AI functionality into applications using the SDK.
Stepping into a larger ecosystem
Microsoft aims to further strengthen its ecosystem by making it easier for Windows users to create AI-powered applications. Although the company does not directly earn revenue from third-party applications, the popularization of AI tools on the platform strengthens Microsoft’s position.
Windows AI Foundry is not just a technological change, but also an open door to the future strategic direction of Windows.
