The Biggest Announcements from Microsoft Build 2025

Hey there, tech friends! Microsoft Build 2025 kicked off yesterday in Seattle, and as always, there's a ton of exciting stuff to cover. This year's theme is crystal clear: AI agents are the future, and Microsoft is going all-in on what they're calling the "open agentic web."

If last year's Build was about laying groundwork, this year is about showing us what's actually possible when AI agents start doing real work. I've dug through the announcements to bring you a breakdown of what I think matters most across Microsoft's ecosystem.

• Microsoft Build 2025 Book of News

Azure AI: Building the Foundation for AI Agents

The most significant shift is the evolution of Azure AI Studio into Azure AI Foundry — Microsoft's comprehensive platform for building and managing AI apps and agents. It's processing a mind-boggling 100 trillion tokens quarterly and supporting over 70,000 customers. The platform includes:

• New models: Microsoft is expanding available models, including integrating xAI’s Grok3 and Grok3 mini into Azure.

• Azure AI Agent Service: Now generally available, this service helps developers build enterprise-grade AI agents that can work together to tackle complex tasks.

• Model Context Protocol (MCP): Microsoft is pushing this open protocol as the "HTTP of AI agents" to enable seamless communication between agents across platforms.

Learn more:

• Building a Digital Workforce with Multi-Agents in Azure AI Foundry Agent Service

• Expand Azure AI Foundry Agent Service with More Knowledge and Action Tools

• Azure AI Foundry: Your AI App and agent factory

• Azure AI Foundry Models: Futureproof Your GenAI Applications

• Announcing General Availability of Azure AI Foundry Agent Service

• What’s new in Azure AI Foundry Fine Tuning

• From diagrams to dialogue: Introducing new multimodal functionality in Azure AI Search

• Introducing agentic retrieval in Azure AI Search

• Announcing Microsoft Entra Agent ID: Secure and manage your AI agents

• Unlock new dimensions of creativity: Gpt-image-1 and Sora

• Announcing the availability of Azure Databricks connector in Azure AI Foundry

• Enterprise-grade controls for AI apps and agents built with Azure AI Foundry and Copilot Studio

Being a believer and supporter in the open web, I’m excited for the introduction of NLWeb. It uses existing open web standards to create natural language interfaces that can be used by people and AI agents. I think it has the potential to improve some of the fundamentals of the web in the long-term and hopefully bring us back from the app-centric nature of today’s internet.

Our goal is to bring the benefits of AI that have transformed the way people search directly to the websites themselves. Just like the introduction of HTML made it easy for almost anyone to create a website, we want NLWeb to make it easy for any web publisher to create an intelligent, natural language experience for their site. And as the agentic web (and economy) continue to grow, NLWeb will empower web publishers to participate on their terms, ensuring their website is ready to interact, transact and be discovered by other agents if they choose.

• Introducing NLWeb: Bringing conversational interfaces directly to the web

The most intriguing new service might be Microsoft Discovery - an enterprise platform designed to accelerate research and development using AI. It combines specialized AI agents with a knowledge engine to help scientists formulate hypotheses, run simulations, and analyze results faster than ever.

Learn more: Transforming R&D with agentic AI: Introducing Microsoft Discovery

GitHub: AI Transforms How We Build Software

GitHub's announcements center around what they're calling "agentic DevOps" - using AI agents to transform the entire software development lifecycle. The headline feature is definitely the GitHub Copilot coding agent, which can now take on tasks ranging from refactoring code to implementing new features, either autonomously or in collaboration with other agents.

In a surprising move that shows Microsoft's commitment to open source, they announced they're open-sourcing GitHub Copilot in VS Code. This means the AI capabilities from GitHub Copilot extensions will become part of the VS Code open-source repository.

There are also new app modernization capabilities aimed specifically at Java and .NET developers, helping them update, upgrade, and modernize applications more quickly.

Learn more:

• Agentic DevOps: Evolving software development with GitHub Copilot and Microsoft Azure

• GitHub Copilot: Meet the new coding agent

• VS Code: Open Source AI Editor

• Reimagining App Modernization for the Era of AI

• Accelerate Your .NET Upgrades with GitHub Copilot

• Public Preview: GitHub Copilot App Modernization for Java

Microsoft 365 & Copilot: Custom AI That Knows Your Business

Microsoft is making Copilot much more customizable and collaborative with these key announcements:

• Microsoft 365 Copilot Tuning: This new low-code capability lets organizations train AI models using their company data and processes without needing data scientists. For example, a law firm could create an agent that generates documents matching their expertise and style.

• Multi-agent orchestration: Now agents can work together as a team! Instead of having one agent try to do everything, you can have specialized agents handling different parts of a complex workflow.

• Copilot Wave 2 Spring Release: This update is now generally available, including an updated Microsoft 365 Copilot app, a new "Create" experience, and Copilot Notebooks.

• Researcher and Analyst agents: These specialized agents for knowledge work are now rolling out globally through the Frontier program.

Learn more:

• What’s new in Microsoft 365 Copilot | May 2025

• Introducing Microsoft 365 Copilot Tuning, multi-agent orchestration, and more

• Build 2025: Agents in Microsoft 365 announcements

• The Microsoft 365 Copilot app: Built for the new way of working

Databases: SQL Server Gets Its Biggest Update in a Decade

Database fans have plenty to be excited about with the preview of SQL Server 2025. Microsoft is calling this the most significant release in the past decade, with built-in AI capabilities that let you:

• Boost search intelligence with advanced semantic search

• Process and manage data more efficiently with native JSON support and built-in REST API

• Improve security with Microsoft Entra ID managed identities through Azure Arc

• Achieve zero-ETL analytics by replicating data to OneLake with Fabric database mirroring

There's also the generally available SQL Server Management Studio 21, based on Visual Studio 2022 with 64-bit support, and a preview of Copilot in SQL Server Management Studio 21 to streamline SQL development.

Learn more:

• Building Intelligent AI Apps with Microsoft Databases

• SQL Server 2025 - AI ready enterprise database from ground to cloud

• SQL Server Management Studio (SSMS) 21 is now generally available

• General Availability of SQL Server Migration Component in SSMS 21

• Announcing the General Availability of JSON data type and JSON aggregates

• New Generative AI Features in Azure Database for PostgreSQL

• Upgrade performance, availability and security with new features in Azure Database for PostgreSQL

• Boost Query Performance with Global Secondary Indexes in Azure Cosmos DB

• Azure AI Foundry Connection for Azure Cosmos DB and BYO Thread Storage in Azure AI Agent Service

• Get to insights faster with SaaS databases and “chat with your data”

• Digital twin builder in Microsoft Fabric Real-Time Intelligence – Revolutionizing digital twin creation and management

Copilot Studio: Better Security and Multi-Agent Workflows

Copilot Studio is getting some serious upgrades focused on security and the ability to coordinate multiple agents:

• Multi-agent orchestration: Now in preview, this allows agents to delegate tasks to one another across Microsoft 365, Azure AI, and Microsoft Fabric.

• Security enhancements: A comprehensive set of new features including advanced connector policies, network isolation, and sensitive data masking.

• Code interpreter: This new feature allows agents to write and run Python code to perform complex tasks like analytics and data visualization.

• Operational database for agents: Powered by Dataverse and optimized for speed and real-time interactions.

Learn more:

• Announcing managed security enhancements for Microsoft Copilot Studio

• Multi-agent orchestration, maker controls, and more: Microsoft Copilot Studio announcements

• Reimagining human-agent collaboration for a new era of app development with Microsoft Power Apps

Azure Announcements beyond the headlines

There were also a number of Azure announcements that didn’t make big headlines, but are worth a mention here:

• Announcing Agent Loop for Azure Logic Apps: An interesting new way to easily embed AI decision making in business process workflows.

• Azure App Service Premium v4 plan is now in public preview: Based on the latest AMD-based virtual machines and NVMe storage, the new v4 App Service plans offer better performance per dollar than earlier options.

• New integrations between Azure Logic Apps and AI Agents in Foundry: You can new use Logic Apps workflows as tools within the Azure Foundry AI Agent Service and use the AI Agent Service connector to build workflows that can trigger agents based on events across hundreds of applications.

• Announcement: Logic Apps connectors in Azure AI Search for Integrated Vectorization: This gives you the ability to ingest unstructured documents from a variety of systems, including SharePoint, Amazon S3, Dropbox and more, into your vector index using the low-code experience of Logic Apps.

• Announcing the open Public Preview of the Premium v2 tier of Azure API Management: The Premium v2 tier uses a simplified architecture that makes private networking easier and more secure.

• Introducing Microsoft Planetary Computer Pro — Now in Public Preview: A fully managed geospatial data platform designed to ingest, catalog, store, process, and disseminate large volumes of private geospatial data in Azure.

• Expose REST APIs as MCP servers with Azure API Management and API Center (now in preview): With this new capability, you can expose your existing REST APIs in Azure API Management as MCP servers—without rebuilding or rehosting them. Also, Azure API Center now serve as a single, enterprise-grade system of record for managing MCP endpoints.

• Build 2025: Announcing Key Improvements in Azure Database for MySQL: New updates to Azure Database for MySQL designed to improve failover performance, operational visibility, and version support for developers and DBAs.

• Introducing Azure SRE Agent: SRE Agent leverages the reasoning capabilities of LLMs to identify the logs and metrics necessary for rapid root cause analysis and issue mitigation. Its advanced AI capabilities transform incident and infrastructure management in Azure, freeing engineers to focus on more meaningful work.

• What's new in Azure Container Apps at Build'25: Highlighting new serverless GPU support, advanced networking capabilities, and improved observability features.

• Announcing Azure AI Language new features to accelerate your agent development: A number of new releases to support AI agent use cases and improve privacy and compliance.

• Build secure, flexible, AI-enabled applications with Azure Kubernetes Service: New features to improve support for AI workloads, boost security, and improve management for AKS operators.

• Azure Functions – Build 2025 update: New tools to help build AI apps with Azure Functions, general availability for running Azure Functions in Azure Container Apps, improvements to triggers and bindings, and more.

• Azure Managed Redis is now generally available: enterprise-grade performance and flexibility: Announcing the general availability of Azure Managed Redis, a fully managed, first-party service built in partnership with Redis

Windows: The AI Platform for Every Device

Windows is evolving to be the best platform for AI developers and users with these announcements:

• Windows AI Foundry: The evolution of last year's Copilot Runtime, this unified platform helps with fine-tuning, optimizing, and deploying AI models on Windows. It automatically detects hardware capabilities and installs necessary components.

• Windows ML improvements: As the built-in AI inferencing runtime, Windows ML now offers simplified model deployment across CPUs, GPUs and NPUs.

• New AI APIs: The Windows App SDK 1.7.2 includes ready-to-use AI APIs for tasks like text summarization, image description, and OCR.

• Microsoft Store updates: Free developer registration, Web Installer for Win32 apps, and new analytics to help developers expand their reach on Windows.

Learn more:

• Advancing Windows for AI development: New platform capabilities and tools introduced at Build 2025

• An IT pro's guide to Windows at Microsoft Build 2025

• The Windows Subsystem for Linux is now open source

Why This All Matters

What's really striking about Build 2025 is how Microsoft is connecting all these pieces into a cohesive vision. They're not just adding AI features — they're reimagining how we'll interact with computers through autonomous agents that can tackle complex tasks across applications, services, and devices.

The open agentic web concept is particularly interesting. By promoting open protocols like MCP and NLWeb, Microsoft is positioning itself at the center of an ecosystem where AI agents can communicate seamlessly with each other regardless of who made them or where they run.

For those of us who build and use software, this shift toward agentic computing means we'll spend less time on repetitive tasks and more time on creative work. Microsoft CTO Kevin Scott even predicted that within five years, 95% of new code could be AI-generated. That's a bold claim, but after seeing these announcements, it doesn't seem too far-fetched.

What announcements caught your eye? Are you excited about building with these new tools or worried about how quickly things are changing? Let me know in the comments!