Microsoft Databases and Microsoft Material: Your unified and AI-powered information property

At present, I’m thrilled to announce the following era of Microsoft’s databases: SQL Server 2025, Azure Doc DB, Azure Horizon DB, and Material Databases, every redesigned to fulfill the calls for of AI.

As AI reshapes each {industry}, one reality stays fixed: information is now not simply an asset—it’s your aggressive edge. The tempo of AI calls for simple information entry, sooner insights, and the power to iterate with out friction. But many organizations are held again by fragmented information estates and legacy methods. Microsoft Material was designed to fulfill this second—to unify your information, simplify your structure, and speed up your path to changing into an AI-led group.

That mission is gaining traction at outstanding pace. Since Material launched two years in the past, it has grown sooner than some other information and analytics platform within the {industry}. Greater than 28,000 prospects—together with 80% of the Fortune 500—now depend on Material, and its ecosystem continues to increase as companions construct options to resolve essentially the most complicated information challenges.

Microsoft already provides one of many {industry}’s most complete database portfolios, and we’re increasing it even additional—whereas deeply integrating these capabilities into Material. I’m excited to announce the overall availability of SQL Server 2025, Azure DocumentDB, and SQL database and Cosmos DB in Material, together with the preview of our latest addition, Azure HorizonDB. With these new choices, you have a world-class database choice to construct as soon as and deploy on the edge, as platform as a service (PaaS), and even as software program as a service (SaaS). And since our whole portfolio is both Material-connected or Material native, Material serves as a unified hub in your whole information property. Beneath I’ll cowl how these new databases are purpose-built to help your AI initiatives.  

Deploy the following era of Microsoft Databases

Modernize your SQL property with SQL Server 2025, now typically out there

Microsoft has been shaping the SQL panorama for greater than 35 years. Now, with the discharge of SQL Server 2025 into basic availability, we’re introducing the following evolution—one which brings developer‑first AI capabilities on the edge, throughout the acquainted T‑SQL expertise. Smarter search combines superior semantic intelligence with full‑textual content filtering to uncover richer insights from complicated information. AI mannequin administration utilizing mannequin definitions in T-SQL permits seamless integration with widespread AI providers akin to Microsoft Foundry.

Enterprise reliability and safety stay best-in-class. Enhanced question efficiency, optimized locking, and improved failover assist guarantee greater concurrency and uptime for mission‑vital workloads. With strengthened credential administration by way of Microsoft Entra ID by way of Azure Arc, SQL Server 2025 is safe by design. Your information can also be immediately accessible in your AI and analytics in Microsoft OneLake with mirroring for SQL Server 2025 in Material, now additionally typically out there.

SQL Server 2025 is essentially the most important launch for SQL builders in a decade. And the response to our preview has been overwhelming, with 10,000 organizations taking part, 100,000 databases already deployed, and obtain charge two instances greater than SQL Server 2022. If you wish to be a part of all those that’ve already adopted SQL Server 2025, obtain it at present.

Azure DocumentDB: MongoDB-compatible, AI-ready, and constructed for hybrid and multi-cloud

We’re excited to announce Azure DocumentDB, a brand new service constructed on the open-source, MongoDB-compatible DocumentDB normal ruled by the Linux Basis. The primary Azure managed service to help multi-cloud and hybrid NoSQL, Azure DocumentDB can run persistently throughout Azure, on-premises, and different clouds.

Azure DocumentDB provides you the liberty to embrace open supply whereas reaching scale, safety, and ease. It’s AI-ready, with capabilities like vector and hybrid search to ship extra related outcomes. On the spot autoscale meets demand, and unbiased compute and storage scaling retains workloads environment friendly. Safety and availability is normal, with Microsoft Entra ID integration, customer-managed encryption keys, 35-day backups included, and a 99.995% availability service-level settlement (SLA). And shortly, enhanced full-text search will add options like fuzzy matching, proximity queries, and expanded language help, making it even simpler to construct clever, search-driven apps.

Azure DocumentDB is now typically out there, so you’ll be able to attempt it at present. You too can be taught extra about Azure DocumentDB and all of the Azure Database information by studying Shireesh Thota’s, Company Vice President of Azure Databases, announcement weblog.

Azure HorizonDB: PostgreSQL designed in your mission-critical workloads

PostgreSQL has grow to be the spine of recent information options because of its wealthy ecosystem, extensibility, and open supply basis. Microsoft is proud to be the #1 PostgreSQL committer amongst hyperscalers, and we’re constructing on that management with Azure HorizonDB.

Now in early preview, Azure HorizonDB is a totally managed, PostgreSQL-compatible database service, constructed to deal with the dimensions and efficiency required by the trendy enterprise. It goes far past open supply Postgres, with auto-scaling storage as much as 128 TB, scale-out compute as much as 3,072 vCores, announcement weblog.

Speed up app growth with Material SaaS Databases, now typically out there

We’re additionally releasing a brand new class of SaaS databases, each SQL database and Cosmos DB in Material, into basic availability. Knowledge builders now have entry to world-class database engines throughout the identical unified platform that powers analytics, AI, and enterprise intelligence.

Material Databases are designed to streamline your software growth. You possibly can provision them in seconds, they usually don’t require the standard granular configuration or deep database experience. They supply enterprise-grade efficiency, are safe by default with options like cloud authentication, customer-managed keys, and database encryption, and are available natively built-in into the Material platform, even utilizing the identical Material capability models for billing.

With Material databases, builders now have the flexibleness to construct functions grounded in operational, transactional, and analytical information. Collectively, these choices make Material a developer-first information platform that’s streamlined, scalable, and prepared for contemporary information functions.

Study extra by studying Shireesh Thota’s, Company Vice President of Azure Databases, announcement weblog.

All of your databases related into Material

We’re making it simpler than ever to work along with your whole Microsoft database portfolio in Material, supplying you with a single, unified place to handle and use all of your information. Constructing on our present mirroring help for Azure SQL Database and Azure SQL MI, we’re now saying the overall availability of mirroring for Azure Database for PostgreSQL, Azure Cosmos DB, and SQL Server variations 2016–2022 and 2025. With these databases mirrored straight into Material, you’ll be able to eradicate conventional extract, rework, and cargo (ETL) pipelines and make your information immediately prepared for analytics and AI.

Getting your information property prepared for AI with Microsoft Material

Choosing the proper database is important, but it surely’s just the start. The most important alternative lies in driving frontier transformation, the place information turns into the inspiration for an AI-native enterprise. We advocate specializing in three core steps:

• Unifying your information property to eradicate silos and complexity.
• Creating semantic that means so your information is prepared for AI.
• Empowering brokers to behave on insights and rework operations.

On this part, I’ll dive into the most recent enhancements to Microsoft Material that assist you obtain each step of your information journey. This contains expanded interoperability in OneLake with SAP, Salesforce, Azure Databricks, and Snowflake, the introduction of Material IQ—a brand new workload that provides semantic understanding—and enhanced agentic capabilities throughout Material that can assist you construct richer, AI-powered information experiences.

That is the way forward for information, and it’s already inside attain. With Material and our database improvements, Microsoft helps organizations transfer seamlessly from perception to motion—unlocking the total potential of your information and the AI constructed on high of it.

Unify your information property with Microsoft OneLake

Microsoft OneLake unifies all of your information—throughout clouds, on-premises, and past Microsoft—right into a single information lake with zero-ETL capabilities like shortcuts and mirroring. Alongside the extra mirroring sources for Microsoft Databases, we’re additionally introducing the preview of shortcuts to SharePoint and OneDrive. This lets you carry unstructured productiveness information into OneLake with out copying information or constructing ETL pipelines, making it simpler to coach brokers and enrich your structured information.

As soon as related to OneLake, your information turns into simply discoverable within the apps your groups use on daily basis like Energy BI, Groups, Excel, Copilot Studio, and Microsoft Foundry. At present, we’re taking {that a} step additional with native integration with Foundry IQ—the following era of retrieval-augmented era (RAG). Brokers depend on context—Foundry IQ’s data bases ship high-value context to brokers by simplifying entry to a number of information sources and making connections throughout info. You need to use the OneLake data supply in Foundry IQ to attach brokers to multi-cloud sources like AWS S3, on-premises sources, and structured and unstructured information.

See how shortcuts and mirroring unify your information in OneLake and gas the following era of clever brokers in Microsoft Foundry:

Increasing OneLake interoperability with main information platforms

We’re additionally seeing nice momentum with dozens of companions exterior of Microsoft deeply integrating with OneLake, together with ClickHouse, Dremio, Confluent, EON, and lots of extra. And now, we’re thrilled so as to add new, deeper interoperability with SAP, Salesforce, Azure Databricks, and Snowflake.

First, we’re deepening interoperability with the methods organizations depend on most, SAP and Salesforce. With the launch of SAP Enterprise Knowledge Cloud Join for Microsoft Material, prospects can permit bidirectional, zero-copy information sharing between SAP Enterprise Knowledge Cloud (BDC) and Material. On the identical time, we’re working with Salesforce to combine their information into Material utilizing the identical zero-copy strategy, unlocking superior analytics and AI capabilities with out the overhead of conventional ETL.

We’re additionally strengthening interoperability with Azure Databricks and Snowflake so you need to use a single copy of information throughout platforms. By the tip of 2025, Azure Databricks will launch, in preview, the power to natively learn information from OneLake by way of Unity Catalog, enabling seamless entry with out duplication or complicated information motion. Trying forward, Databricks may even add help for writing to and storing information straight in OneLake, permitting full two-way interoperability. Learn extra about this interoperability.

Our collaboration with Snowflake on bidirectional information entry continues as properly. We’re introducing a brand new merchandise in OneLake referred to as a Snowflake Database and a brand new UI in Snowflake—each designed to permit OneLake to be the native storage resolution in your Snowflake information. We’re additionally bringing Snowflake mirroring to basic availability, permitting you to virtualize your exterior Snowflake-managed Iceberg tables in OneLake with shortcuts created and dealt with robotically. Collectively, these improvements allow you to run any Material workload—whether or not analytics, AI, or visualization—straight in your Snowflake-managed Iceberg tables.

Study extra about our Snowflake collaboration by studying our newest joint weblog or by watching the next demo:

Lastly, in shut collaboration with dbt Labs, we’re additionally excited to announce built-in help for his or her {industry} main information transformation functionality. Now in preview, dbt jobs in Microsoft Material allow you to construct, take a look at, and orchestrate dbt workflows in your Material workspaces. Study extra on this weblog.

Create semantic data to gas AI with Material IQ

As Frontier Companies practice brokers on their enterprise information, it’s grow to be clear that high quality and context matter greater than information quantity. Brokers want enterprise context throughout relationships, hierarchies, and that means to show uncooked information into actionable perception. That’s why we’re introducing Material IQ—a brand new workload designed to map your datasets to the real-world entities they symbolize, making a shared semantic construction on high of your information.

The ability of IQ lies in the way it unifies disparate information varieties beneath a single, coherent framework. Constructed upon Energy BI’s industry-leading, wealthy semantic mannequin expertise, IQ brings collectively analytical information, time-series telemetry, and geospatial info, all organized beneath a semantic framework of enterprise entities and their relationships, properties, guidelines, and actions. You possibly can then create operations brokers, a brand new sort of agent in Material, which might use this mannequin to behave as digital staff members, monitoring real-time information sources, figuring out patterns, and taking proactive motion. As an alternative of forcing your groups and even brokers to suppose when it comes to tables and schemas, IQ means that you can align information with how your group operates.

Briefly, Material IQ is designed to mannequin actuality with information, so that each perception, prediction, and motion is grounded in how your group really operates. You possibly can be taught extra about IQ in Yitzhak Kesselman’s, Company Vice President of Messaging and Actual-Time Intelligence, announcement weblog.

As organizations scale their AI initiatives, the power to attach clever brokers with enterprise-grade information is changing into a vital differentiator. Material is making this doable with a set of built-in AI experiences: Copilot in Energy BI helps you ask questions of your information, Material information brokers permit deeper evaluation, and the brand new Material operations brokers allow you to monitor your information property and take motion in actual time. These experiences can be utilized throughout Material or as foundational data sources in industry-leading AI instruments like Microsoft Foundry, Copilot Studio and even Microsoft 365 Copilot to energy smarter, extra data-rich AI experiences.

Past introducing operations brokers as a part of Material IQ, we’re additionally increasing what information brokers and Copilot can do. Together with present integration with Microsoft Foundry and Copilot Studio, Material information brokers can now be embedded straight in Microsoft 365 Copilot. This lets enterprise customers (with the correct permissions) entry trusted data from OneLake and transforms Microsoft 365 from a productiveness suite into an clever insights platform.

They will additionally act as hosted Mannequin Context Protocol (MCP) servers, making it simple to combine with different functions and brokers throughout the AI ecosystem. Lastly, information brokers can now cause throughout each structured and unstructured information. Due to an integration with Azure AI Search, information groups can add their present unstructured information search endpoints as a supply in information brokers. Study extra the Material information agent enhancements by studying the Material AI weblog.

We’re additionally enhancing the standalone expertise for Copilot in Energy BI with a brand new search expertise. Merely describe what you want, and Copilot will find the related report, semantic mannequin, or information agent and floor the correct solutions. This standalone expertise can also be coming to Energy BI cellular so you need to use it on the go.

Check out how one can apply all of those AI experiences collectively seamlessly:

Briefly, we’re redefining what it means to have an AI-powered information property. With information brokers, Copilot in Energy BI, and operations brokers in Material IQ, AI is now woven throughout Material. And with native integration to Microsoft Foundry and Copilot Studio, you’ll be able to simply add Material brokers as constructing blocks to create extra clever, knowledgeable customized brokers.

You can also see extra innovation coming to the Material platform by studying Kim Manis’, Company Vice President of the Material Platform, Material weblog or by trying out the extra technical Material November 2025 Characteristic abstract weblog.

Mark your calendar for FabCon and SQLCon

We’re excited to announce SQLCon 2026, which is able to occur on the identical time and the identical location because the Microsoft Material Group Convention (FabCon), occurring March 16–20, 2026 in Atlanta, Georgia. By uniting the powerhouse SQL and Material communities, we’re giving information professionals in all places a novel alternative to grasp the most recent improvements, share sensible data, and speed up what’s doable with information and AI, multi function highly effective week. Register for both convention and luxuriate in full entry to each, with the flexibleness to combine and match periods, keynotes, and group occasions to suit your pursuits.

For those who’re desirous about seeing these bulletins stay, I encourage you to affix my Ignite session, “Innovation Session: Microsoft Material and Azure Databases – the info property for AI” both in individual or on-line for free of charge. I’ll not solely cowl these main bulletins however present you the way they arrive collectively that can assist you create a unified, clever information basis for AI.

You too can dive deeper into these bulletins and a lot extra by watching the remainder of the breakout periods throughout Azure Knowledge:

Tuesday, November 18