A summary of Microsoft’s latest innovations across Fabric, Azure Data, and developer tools (June 2025)
- Microsoft’s latest Work Trend Index introduces the concept of “agentic organizations”—companies leveraging AI agents to automate tasks, augment teams, or even lead AI-only teams. This shift requires developers to integrate operational, transactional, and analytical data to power these intelligent agents.
- Microsoft Fabric continues to evolve as the unified, AI-powered data platform for organizations. Over 21,000 customers, including 70% of the Fortune 500, are already using Fabric to unify their data and accelerate innovation.
- Fabric Databases now supports Cosmos DB, enabling developers to work with semi-structured data like text, graphs, and emails—critical for training AI agents. Cosmos DB’s scalability and performance make it ideal for generative AI workloads.
- A new Digital Twin Builder in Fabric is now in preview. It allows no-code/low-code creation of virtual models of physical or logical entities (like machines or customer processes) to drive advanced analytics and automation.
- A standalone Copilot experience in Power BI now lets users chat with their data across reports and models. This new feature is embedded in Microsoft 365 apps and Teams, making insights easier to access in daily workflows.
- Fabric Data Agents can now be integrated with Microsoft Copilot Studio, enabling users to interact with enterprise data using natural language and even trigger actions (like sending emails or workflows) across Teams and Microsoft 365.
- Microsoft announced SQL Server 2025 (Public Preview), reimagined to support vector search and AI model integration, allowing real-time contextual reasoning directly within the database.
- The new PostgreSQL extension for VS Code integrates GitHub Copilot, allowing developers to write queries, design schemas, and troubleshoot faster using AI—available for Azure, on-prem, or Docker.
- DiskANN, a high-performance vector indexing engine, is now available in Azure PostgreSQL, enabling faster and more scalable AI applications than vector.
- Microsoft introduced semantic operators in Azure PostgreSQL to reason over data using large language models (LLMs), helping extract deeper meaning from relational data.
- Azure Cosmos DB now supports conversation thread storage and integrates with Azure AI Foundry, helping agents maintain memory and context in real-time. Data from Cosmos DB can be directly used by AI agents via Foundry.
- New Azure Databricks + Foundry integration lets AI agents use Databricks jobs and knowledge, enhancing contextual data access and intelligent responses.
- A new SAP Databricks on Azure experience is launching in Q3 2025. It will support SAP Business Data Cloud with full Azure-native integration, allowing analytics on SAP data at scale.
- Mirroring now connects Azure Databricks Unity Catalog tables with Fabric One Lake—enabling real-time access and syncing of Databricks data into Fabric for unified analytics.
- These updates reflect Microsoft’s strategy to break down silos and unify all data—structured or unstructured, cloud or on-prem—to empower developers building AI-powered applications and agents.