Windows AI Foundry will support the full development lifecycle across various hardware, enabling developers to build applications that run seamlessly on all devices.

We created AI Foundry to help people build their own AIs, providing the necessary tooling and expertise to turn their ideas into AI services.

The Foundry agent service allows for the creation of declarative agents with minimal code, supporting complex workflows and multi-agent orchestration.

Microsoft aims to create an agentic web where tools like Microsoft 365 Copilot and Foundry work together seamlessly.

The output of our AI Foundry is a Neuro Micro NVIDIA Inference Microservice (NIM) that users can run anywhere, including on-premises.

The integration of Cosmos DB into Foundry allows agents to store and retrieve conversational history and supports various application needs, enhancing the capabilities of AI applications.

Fine-tuning Copilot with proprietary data will give companies a competitive edge over generic AI solutions. The sustainable advantage lies in obtaining new samples to apply reasoning models with your data, enabling reinforcement learning in real-world applications.

Developers can leverage the latest Gemini and Gemma models in AI Studio to create robust AI experiences and deploy them easily.

We're open sourcing GitHub Copilot in VS Code, integrating AI-powered capabilities directly into the core of VS Code.