At Microsoft Build 2025, we introduced Azure AI Foundry resource, Azure AI Foundry API, and supporting tools to streamline the end-to-end development lifecycle of AI agents and applications. 

These capabilities are designed to help developers accelerate time-to-market; support production-scale workloads with scale and central governance; and support administrators with a self-serve capability to enable their teams’ experimentation with AI in a controlled environment.

The Azure AI Foundry resource type unifies agents, models and tools under a single management grouping, equipped with built-in enterprise-readiness capabilities — such as tracing & monitoring, agent and model-specific evaluation capabilities, and customizable enterprise setup configurations tailored to your organizational policies like using your own virtual networks. This launch represents our commitment to providing organizations with a consistent, efficient and centrally governable environment for building and operating the AI agents and applications of today, and tomorrow.

New platform capabilities

The new Foundry resource type evolves our vision for Azure AI Foundry as a unified Azure platform-as-a-service offering, enabling developers to focus on building applications rather than managing infrastructure, while taking advantage of native Azure platform capabilities like Azure Data and Microsoft Defender. Previously, Azure AI Foundry portal’s capabilities required the management of multiple Azure resources and SDKs to build an end-to-end application.

New capabilities include:

• Foundry resource type enables administrators with a consistent way of managing security and access to Agents, Models, Projects, and Azure tooling Integration. With this change, Azure Role Based Access Control, Networking and Policies are administered under a single Azure resource provider namespace, for streamlined management. ‘Azure AI Foundry’ is a renaming of the former ‘Azure AI Services’ resource type, with access to new capabilities. While Azure AI Foundry still supports bring-your-own Azure resources, we now default to a fully Microsoft-managed experience, making it faster and easier to get started.


• Foundry projects are folders that enable developers to independently create new environments for exploring new ideas and building prototypes, while managing data in isolation. Projects are child resources; they may be assigned their own admin controls but by default share common settings such as networking or connected resource access from their parent resource. This principle aims to take IT admins out of the day-to-day loop once security and governance are established at the resource level, enabling developers to self-serve confidently within their projects.


• Azure AI Foundry API is designed from the ground up, to build and evaluate API-first agentic applications, and lets you work across model providers agnostically with a consistent contract. 


• Azure AI Foundry SDK wraps the Foundry API making it easy to integrate capabilities into code whether your application is built in Python, C#, JavaScript/TypeScript or Java.


• Azure AI Foundry for VS Code Extension complements your workflow with capabilities to help you explore models, and develop agents and is now supported with the new Foundry project type.


• New built-in RBAC roles provide up-to-date role definitions to help admins differentiate access between Administrator, Project Manager and Project users. Foundry RBAC actions follow strict control- and data plane separation, making it easier to implement the principle of least privilege.

Why we built these new platform capabilities

If you are already building with Azure AI Foundry — these capabilities are meant to simplify platform management, enhance workflows that span multiple models and tools, and reinforce governance capabilities, as we see AI workloads grow more complex.

The emergence of generative AI fundamentally changed how customers build AI solutions, requiring capabilities that span multiple traditional domains. We launched Azure AI Foundry to provide a comprehensive toolkit for exploring, building and evaluating this new wave of GenAI solutions. Initially, this experience was backed by two core Azure services — Azure AI Services for accessing models including those from OpenAI, and Azure Machine Learning’s hub, to access tools for orchestration and customization.

With the emergence of AI agents composing models and tools; and production workloads demanding the enforcement of central governance across those, we are investing to bring the management of agents, models and their tooling integration layer together to best serve these workload’s requirements.

The Azure AI Foundry resource and Foundry API are purposefully designed to unify and simplify the composition and management of core building blocks of AI applications:

• Models


• Agents & their tools


• Observability, Security, and Trust

In this new era of AI, there is no one-size-fits-all approach to building AI agents and applications. That’s why we designed the new platform as a comprehensive AI factory with modular, extensible, and interoperable components.

Foundry Project vs Hub-Based Project

Going forward, new agents and model-centric capabilities will only land on the new Foundry project type. This includes access to Foundry Agent Service in GA and Foundry API.

While we are transitioning to Azure AI Foundry as a managed platform service, hub-based project type remains accessible in Azure AI Foundry portal for GenAI capabilities that are not yet supported by the new resource type. Hub-based projects will continue to support use cases for custom model training in Azure Machine Learning Studio, CLI and SDK.

For a full overview of capabilities supported by each project type, see this support matrix.

Azure AI Foundry Agent Service  

The Azure AI Foundry Agent Service experience, now generally available, is powered by the new Foundry project. Existing customers exploring the GA experience will need the new AI Foundry resource.

All new investments in the Azure AI Foundry Agent Service are focused on the Foundry project experience. Foundry projects act as secure units of isolation and collaboration — agents within a project share:

• File storage


• Thread storage (i.e. conversation history)


• Search indexes

You can also bring your own Azure resources (e.g., storage, bring-your-own virtual network) to support compliance and control over sensitive data.

Start Building with Foundry

Azure AI Foundry is your foundation for scalable, secure, and production-grade AI development. Whether you’re building your first agent or deploying a multi-agent workforce at Scale, Azure AI Foundry is ready for what’s next.