Microsoft Unveils Windows AI Foundry: A New Era for AI Development on PCs

Microsoft Introduces Windows AI Foundry

Microsoft has unveiled Windows AI Foundry, a significant upgrade to its AI development tools for Windows, marking a new chapter in the company's AI strategy. Announced at the Build 2025 conference, this platform replaces and expands upon the Windows Copilot Runtime service launched in May 2024 1

1

2

2

.

A Unified Platform for Local AI Development

Windows AI Foundry is described as a "unified platform for local AI development," designed to simplify the process of fine-tuning, optimizing, and deploying AI models for Windows applications 1

1

. This move aligns with Microsoft's vision of AI as a major profit driver, building on its $13 billion annualized AI revenue reported in January 2025 1

1

.

Key Features of Windows AI Foundry

1. Automatic Hardware Detection: The platform can automatically detect a Windows machine's hardware and fetch necessary software components to run specific AI models 1

   1

   3

   3

   .

2. Foundry Local: This new service brings the power of AI models directly to client devices, supporting both Windows and macOS. It includes modules to run AI models on-device, leveraging the ONXX Runtime 1

   1

   4

   4

   .

3. Command-Line Interface Support: Developers can use prompts like "Foundry model list" and "Foundry model run" to interact with models running on a local server 1

   1

   .

4. Multi-Model Support: Windows AI Foundry includes not only Microsoft's models but also open-source models from other providers like Nvidia 2

   2

   .

5. Ready-to-Use AI APIs: The platform offers APIs for key language and vision tasks, such as text intelligence, image description, and object erase 2

   2

   .

Advanced Capabilities for Developers

Windows AI Foundry introduces several advanced features for developers:

1. LoRA (Low-Rank-Adaption): This feature allows for fine-tuning of Microsoft's inbox SLM, Phi Silica, with custom data 2

   2

   4

   4

   .

2. Semantic Search and RAG: New APIs enable semantic search and knowledge retrieval for natural language search and retrieval-augmented generation scenarios 2

   2

   .

3. Windows ML: An evolution of the DirectML runtime, Windows ML acts as the built-in AI inference runtime to streamline model deployment across various hardware types 4

   4

   .

Impact on PC Computing

Microsoft sees this shift towards running advanced AI locally on devices as a major turning point in computing. Stevie Bathiche, Microsoft Technical Fellow, describes AI agents as "the new unit of interaction" and a "new unit of programming" 3

3

.

Copilot+ PCs and Future Outlook

Microsoft is introducing Copilot+ PCs, new Windows laptops and desktops built to handle AI tasks directly on the device. The company expects that within a few years, most new Windows machines will fall into this category, making advanced AI tools an everyday part of the PC experience 3

3

.

Integration with Azure AI Foundry

For cloud-focused developers, Microsoft is also updating Azure AI Foundry, a comprehensive platform for designing, building, customizing, and managing AI applications and agents in the cloud 4

4

.

This strategic move by Microsoft aims to position Windows at the center of the AI revolution, much like it did during the early days of the web. By providing powerful tools for local AI development, Microsoft is fostering an ecosystem that could give it an edge over rivals in the rapidly evolving AI landscape.