Microsoft is including a number of initiatives that aim to help address organization data complexity, as well as an additional database option addition to the Microsoft Fabric collection.

Practically speaking, database years often had close connections to quantify and storage. That created a wide range of scaling and data bin problems for businesses. Microsoft Fabric was first introduced as a method to beat that challenge in 2023. The fundamental principle of Microsoft Fabric is to serve as a common data level between Microsoft’s data and analytics resources. Microsoft Fabric expanded with assistance for the Azure SQL contextual collection system in November 2024.

Microsoft has a number of different database systems, just like its competitors at Google and Amazon. Although Azure SQL is commonly used, CosmosDB is a more potent AI database system. Microsoft is announcing that CosmosDB is suddenly coming to Microsoft Fabric at the Build 2025 event today. CosmosDB is one of the most crucial data currently being used by AI because it serves as the foundation for OpenAI’s ChatGPT company. Additionally, CosmosDB’s connectivity with Azure AI Foundry, which gives agentic AI more immediate access to data, is helping.

Additionally, the PowerBI business intelligence system supports a number of more information updates, including one for Microsoft Copilot. The DiskANN ( Disk Approximate Nearest Neighbor ) vector index is being open-sourced while the SQL Server 2025 database is being previewed.

These advancements address the connectivity difficulty that enterprise data teams encounter when developing AI applications. Eliminating the data separation that impedes business AI initiatives is a key priority.

Arun Ulag, Corporate Vice President for Azure Data at Microsoft, told VentureBeat,” The information I constantly receive is, please consolidate, &nbsp, I’m Chief Information Officer, I don’t want to be the Chief Integration Officer helping convert AI into my aggressive advantage.”

Fabric reduces information silos and speeds up business AI.

Microsoft Fabric, the business’s unified information system, keeps on its fast growth path by bringing recently unrelated products along in a consolidated ecosystem.

According to Ulag,” We’re bringing all of our goods up and unifying them into one item, which is Microsoft Fabric.” You can think of Fabric in some ways as nearly like what we did with Office 30 years ago.

This approach has proven to be effective for businesses. According to Ullag, Microsoft Fabric currently has over 21, 000 businesses as paying clients in the world, including 70 % of the Fortune 500. &nbsp,

He claimed that “it’s growing pretty, very quickly.”

Eliminates the costs associated with NoSQL network with CosmosDB in Fabric.

CosmosDB, a Microsoft’s NoSQL report database that powers numerous well-known AI applications, is the title addition to Fabric.

CosmosDB is, by far, frequently becoming the collection of selection for the world’s AI tasks, according to Ulag. The e-commerce business at Walmart also runs on CosmosDB, according to the statement” ChatGPT itself is built on CosmosDB.”

Microsoft makes it easier for businesses to install NoSQL databases without managing a complicated system by integrating CosmosDB with Fabric. Maintaining efficiency without overhead is a key concern of having a fragmented determine and storage approach.

Microsoft has implemented incredibly precise professional measures to maintain performance through a cutting-edge caching system.

We maintain a very efficient memory inside Fabric that can handle all the quick revisions CosmosDB can perform, according to Ulag. We have a very quick synchronization process that is fully open to the customer, where the information is replicated into OneLake almost instantly.

This method eliminates network management tasks while delivering the second response times needed for AI applications.

Why the success of Fabric depends in large part on open source data types.

OneLake tech stores the data while Microsoft uses the Fabric method to connect all of its data products.

A unified information cove that supports a variety of information types and forms from SQL, NoSQL, and unstructured data is a lot of work. Microsoft is tackling this issue with an empty source view.

According to Ulag,” Microsoft has fully embraced open source data types, so anything in Fabric, irrespective of workload, is always in Apache Parquet and Delta Lake by default.” With all of the information stored in a global SaaS data river, which is OneLake in open source data formats, “it’s actually a unified solution, with the consolidated infrastructure and a consolidated business design.”

This improvement makes it possible for all Textile services, from SQL to Power BI to CosmosDB, to access the same core data without transformation or duplication, thereby reducing the traditionally slow performance issue that comes with available formats.

Enterprise-grade matrix search is made available with the DiskANN open source transfer.

Microsoft contributes its own code as well, but it doesn’t really use open source for information formats.

Microsoft is announcing at Build that it is empty purchasing the DiskANN matrix search technologies. Microsoft’s determination to start source DiskANN makes a significant contribution to the AI ecology, enabling all designers to access enterprise-grade vector search features.

DiskANN, which was first developed for Microsoft Research and later added to CosmosDB and Fabric, is a very, very powerful vector capacity, according to Ulag.

DiskANN implements ANN search algorithms that are tailored for disk-based procedures, making it ideal for large-scale vectors databases that use more storage resources. Microsoft gives developers the similar high-performance variable search that is used by ChatGPT and other top AI applications by empty purchasing DiskANN. This addresses one of the biggest issues with creating retrieval-augmented generation (RAG ) systems, where it is crucial to quickly locate semantically similar content for grounding AI responses in enterprise data.

According to Ulag,” We’re allowing everyone to enjoy the advantages of the matrix keep that we’re using internally.”

Why do business info leaders care?

These announcements make it possible for more complex applications that seamlessly incorporate various data types for enterprises that are in the forefront of AI adoption.

The challenges and difficulty of dealing with data silos are not just related to different formats, but also to various locations. No additional hyperscaler is currently addressing that problem in a way that Microsoft Fabric’s continued evolution does. &nbsp,

The primary concern and dedication to open source requirements is also crucial for businesses because it removes some of the lock-in risk that might exist if the data were encased in specialized formats.

Microsoft’s unified technique removes a major challenge to innovation as businesses increasingly compete for AI capabilities. Organizations who embrace this integration may shift their focus away from maintaining difficult data pipelines and toward developing AI applications that deliver tangible business value, potentially outperforming their competitors who are still struggling with divided architectures.

With VB Daily, you get regular perspectives on business use cases.

VB Daily has you covered if you want to please your employer. In order to gain the most from your insight on how companies are using conceptual AI, from regulatory changes to functional deployments, we provide the inside scoop on what businesses are doing.

Learn our privacy statement.

Thank you for your subscription. View more VB newsletters around.

An failure took place.