Formerly known as Azure AI Services or Azure Cognitive Services is a unified collection of prebuilt AI capabilities within the Microsoft Foundry platform
Hello Meenakshi Sankar,
Welcome to Microsoft Q&A and Thank you for reaching out.
I understand that you’re evaluating the Classic vs. New experience in Azure Foundry. Below is a consolidated explanation covering portal differences, capabilities, recommendations, and lifecycle status.
Azure Foundry Classic vs. Azure Foundry (New)
1.Portal & Project Structure
Foundry Classic Portal
Labeled as “Microsoft Foundry (classic)” in the portal.
Supports:
- Hub-based projects (existing ML hubs, Azure OpenAI resources, etc.)
- Foundry projects
Allows switching to the New experience via the header toggle.
Requires manual management of backend Azure resources when building solutions:
- Storage accounts
- Key Vault
- Separate ML endpoints
Follows a more traditional Azure resource-centric setup.
Classic reflects the original architecture and project model.
Foundry (New) Portal
Focused purely on Foundry-native projects (no hub concept).
Does not expose additional infrastructure primitives.
Provides a streamlined, unified UX for:
- Agents
- Models
- Tools
Only Foundry-native projects appear in this experience.
It abstracts much of the infrastructure wiring and simplifies development workflows.
Capability Differences
Foundry New introduces more modern platform capabilities, including:
Multi-agent orchestration
Private and public tool catalogs
Enhanced memory handling
Integrated knowledge base workflows
Real-time observability and tracing
Centralized asset management
Faster load times and simplified navigation
Classic remains functional but has a more static and infrastructure-heavy setup compared to the New experience.
Recommended Platform for Future Development
Microsoft’s strategic investments are focused on Foundry (New).
New features are introduced there first.
SDKs and APIs are aligned with the New experience.
It is designed for forward-looking, enterprise-scale AI development.
For any greenfield or future projects, Foundry (New) is the recommended platform.
Retirement Timeline for Foundry Classic
There is no officially announced deprecation date for Foundry Classic.
Existing workloads can continue running without immediate impact.
However, feature development is centered on the New experience.
While Classic is still supported, it is not the strategic direction moving forward. Planning new projects on Foundry (New) ensures access to ongoing enhancements.
Foundry Classic: Original hub-based architecture; supported but legacy direction.
Foundry (New): Modern, streamlined, feature-rich platform; recommended for new development.
Retirement: No announced timeline, but innovation focus is on the New experience.
Please refer this
- What is Microsoft Foundry? (covers Classic vs New at a glance) https://learn.microsoft.com/azure/foundry/what-is-foundry?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider
- Choose an Azure resource type for Foundry (classic) https://learn.microsoft.com/azure/foundry-classic/concepts/resource-types?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider
- Migrate from hub-based to Foundry projects (classic) https://learn.microsoft.com/azure/foundry-classic/how-to/migrate-project?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider
I Hope this helps. Do let me know if you have any further queries.
If this answers your query, could you please take a moment to retake the survey on the response by accepting this response? Your feedback is greatly appreciated.
Thank you!